diff options
author | Vlastimil Holer <vlastimil.holer@gmail.com> | 2013-09-05 13:11:09 +0200 |
---|---|---|
committer | Vlastimil Holer <vlastimil.holer@gmail.com> | 2013-09-05 13:11:09 +0200 |
commit | 744c779182cba32314f8435660a61c2711cb9f54 (patch) | |
tree | 7871342bf0b122217b51493286bac982313b48da /cloudinit | |
parent | 8a2a88e0bb4520eabe99b6686413a548f3d59652 (diff) | |
parent | 1d27cd75eaaeef7b72f3be77de24da815c82a825 (diff) | |
download | vyos-cloud-init-744c779182cba32314f8435660a61c2711cb9f54.tar.gz vyos-cloud-init-744c779182cba32314f8435660a61c2711cb9f54.zip |
Merged trunk lp:cloud-init
Diffstat (limited to 'cloudinit')
53 files changed, 3171 insertions, 863 deletions
diff --git a/cloudinit/config/cc_apt_configure.py b/cloudinit/config/cc_apt_configure.py index f8664160..5a407016 100644 --- a/cloudinit/config/cc_apt_configure.py +++ b/cloudinit/config/cc_apt_configure.py @@ -27,7 +27,8 @@ from cloudinit import util distros = ['ubuntu', 'debian'] PROXY_TPL = "Acquire::HTTP::Proxy \"%s\";\n" -PROXY_FN = "/etc/apt/apt.conf.d/95cloud-init-proxy" +APT_CONFIG_FN = "/etc/apt/apt.conf.d/94cloud-init-config" +APT_PROXY_FN = "/etc/apt/apt.conf.d/95cloud-init-proxy" # A temporary shell program to get a given gpg key # from a given keyserver @@ -67,18 +68,10 @@ def handle(name, cfg, cloud, log, _args): "security": "security.ubuntu.com/ubuntu"}) rename_apt_lists(old_mirrors, mirrors) - # Set up any apt proxy - proxy = cfg.get("apt_proxy", None) - proxy_filename = PROXY_FN - if proxy: - try: - # See man 'apt.conf' - contents = PROXY_TPL % (proxy) - util.write_file(proxy_filename, contents) - except Exception as e: - util.logexc(log, "Failed to write proxy to %s", proxy_filename) - elif os.path.isfile(proxy_filename): - util.del_file(proxy_filename) + try: + apply_apt_config(cfg, APT_PROXY_FN, APT_CONFIG_FN) + except Exception as e: + log.warn("failed to proxy or apt config info: %s", e) # Process 'apt_sources' if 'apt_sources' in cfg: @@ -140,10 +133,13 @@ def get_release(): def generate_sources_list(codename, mirrors, cloud, log): - template_fn = cloud.get_template_filename('sources.list') + template_fn = cloud.get_template_filename('sources.list.%s' % + (cloud.distro.name)) if not template_fn: - log.warn("No template found, not rendering /etc/apt/sources.list") - return + template_fn = cloud.get_template_filename('sources.list') + if not template_fn: + log.warn("No template found, not rendering /etc/apt/sources.list") + return params = {'codename': codename} for k in mirrors: @@ -253,3 +249,22 @@ def find_apt_mirror_info(cloud, cfg): mirror_info.update({'primary': mirror}) return mirror_info + + +def apply_apt_config(cfg, proxy_fname, config_fname): + # Set up any apt proxy + cfgs = (('apt_proxy', 'Acquire::HTTP::Proxy "%s";'), + ('apt_http_proxy', 'Acquire::HTTP::Proxy "%s";'), + ('apt_ftp_proxy', 'Acquire::FTP::Proxy "%s";'), + ('apt_https_proxy', 'Acquire::HTTPS::Proxy "%s";')) + + proxies = [fmt % cfg.get(name) for (name, fmt) in cfgs if cfg.get(name)] + if len(proxies): + util.write_file(proxy_fname, '\n'.join(proxies) + '\n') + elif os.path.isfile(proxy_fname): + util.del_file(proxy_fname) + + if cfg.get('apt_config', None): + util.write_file(config_fname, cfg.get('apt_config')) + elif os.path.isfile(config_fname): + util.del_file(config_fname) diff --git a/cloudinit/config/cc_bootcmd.py b/cloudinit/config/cc_bootcmd.py index 896cb4d0..3ac22967 100644 --- a/cloudinit/config/cc_bootcmd.py +++ b/cloudinit/config/cc_bootcmd.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2009-2011 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # # Author: Scott Moser <scott.moser@canonical.com> # Author: Juerg Haefliger <juerg.haefliger@hp.com> @@ -50,6 +50,5 @@ def handle(name, cfg, cloud, log, _args): cmd = ['/bin/sh', tmpf.name] util.subp(cmd, env=env, capture=False) except: - util.logexc(log, - ("Failed to run bootcmd module %s"), name) + util.logexc(log, "Failed to run bootcmd module %s", name) raise diff --git a/cloudinit/config/cc_chef.py b/cloudinit/config/cc_chef.py index 607f789e..727769cd 100644 --- a/cloudinit/config/cc_chef.py +++ b/cloudinit/config/cc_chef.py @@ -110,7 +110,7 @@ def handle(name, cfg, cloud, log, _args): with util.tempdir() as tmpd: # use tmpd over tmpfile to avoid 'Text file busy' on execute tmpf = "%s/chef-omnibus-install" % tmpd - util.write_file(tmpf, content, mode=0700) + util.write_file(tmpf, str(content), mode=0700) util.subp([tmpf], capture=False) else: log.warn("Unknown chef install type %s", install_type) diff --git a/cloudinit/config/cc_growpart.py b/cloudinit/config/cc_growpart.py new file mode 100644 index 00000000..2d54aabf --- /dev/null +++ b/cloudinit/config/cc_growpart.py @@ -0,0 +1,277 @@ +# vi: ts=4 expandtab +# +# Copyright (C) 2011 Canonical Ltd. +# Copyright (C) 2013 Hewlett-Packard Development Company, L.P. +# +# Author: Scott Moser <scott.moser@canonical.com> +# Author: Juerg Haefliger <juerg.haefliger@hp.com> +# +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License version 3, as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. + +import os +import os.path +import re +import stat + +from cloudinit import log as logging +from cloudinit.settings import PER_ALWAYS +from cloudinit import util + +frequency = PER_ALWAYS + +DEFAULT_CONFIG = { + 'mode': 'auto', + 'devices': ['/'], +} + + +def enum(**enums): + return type('Enum', (), enums) + + +RESIZE = enum(SKIPPED="SKIPPED", CHANGED="CHANGED", NOCHANGE="NOCHANGE", + FAILED="FAILED") + +LOG = logging.getLogger(__name__) + + +def resizer_factory(mode): + resize_class = None + if mode == "auto": + for (_name, resizer) in RESIZERS: + cur = resizer() + if cur.available(): + resize_class = cur + break + + if not resize_class: + raise ValueError("No resizers available") + + else: + mmap = {} + for (k, v) in RESIZERS: + mmap[k] = v + + if mode not in mmap: + raise TypeError("unknown resize mode %s" % mode) + + mclass = mmap[mode]() + if mclass.available(): + resize_class = mclass + + if not resize_class: + raise ValueError("mode %s not available" % mode) + + return resize_class + + +class ResizeFailedException(Exception): + pass + + +class ResizeParted(object): + def available(self): + myenv = os.environ.copy() + myenv['LANG'] = 'C' + + try: + (out, _err) = util.subp(["parted", "--help"], env=myenv) + if re.search(r"COMMAND.*resizepart\s+", out, re.DOTALL): + return True + + except util.ProcessExecutionError: + pass + return False + + def resize(self, diskdev, partnum, partdev): + before = get_size(partdev) + try: + util.subp(["parted", diskdev, "resizepart", partnum]) + except util.ProcessExecutionError as e: + raise ResizeFailedException(e) + + return (before, get_size(partdev)) + + +class ResizeGrowPart(object): + def available(self): + myenv = os.environ.copy() + myenv['LANG'] = 'C' + + try: + (out, _err) = util.subp(["growpart", "--help"], env=myenv) + if re.search(r"--update\s+", out, re.DOTALL): + return True + + except util.ProcessExecutionError: + pass + return False + + def resize(self, diskdev, partnum, partdev): + before = get_size(partdev) + try: + util.subp(["growpart", '--dry-run', diskdev, partnum]) + except util.ProcessExecutionError as e: + if e.exit_code != 1: + util.logexc(LOG, "Failed growpart --dry-run for (%s, %s)", + diskdev, partnum) + raise ResizeFailedException(e) + return (before, before) + + try: + util.subp(["growpart", diskdev, partnum]) + except util.ProcessExecutionError as e: + util.logexc(LOG, "Failed: growpart %s %s", diskdev, partnum) + raise ResizeFailedException(e) + + return (before, get_size(partdev)) + + +def get_size(filename): + fd = os.open(filename, os.O_RDONLY) + try: + return os.lseek(fd, 0, os.SEEK_END) + finally: + os.close(fd) + + +def device_part_info(devpath): + # convert an entry in /dev/ to parent disk and partition number + + # input of /dev/vdb or /dev/disk/by-label/foo + # rpath is hopefully a real-ish path in /dev (vda, sdb..) + rpath = os.path.realpath(devpath) + + bname = os.path.basename(rpath) + syspath = "/sys/class/block/%s" % bname + + if not os.path.exists(syspath): + raise ValueError("%s had no syspath (%s)" % (devpath, syspath)) + + ptpath = os.path.join(syspath, "partition") + if not os.path.exists(ptpath): + raise TypeError("%s not a partition" % devpath) + + ptnum = util.load_file(ptpath).rstrip() + + # for a partition, real syspath is something like: + # /sys/devices/pci0000:00/0000:00:04.0/virtio1/block/vda/vda1 + rsyspath = os.path.realpath(syspath) + disksyspath = os.path.dirname(rsyspath) + + diskmajmin = util.load_file(os.path.join(disksyspath, "dev")).rstrip() + diskdevpath = os.path.realpath("/dev/block/%s" % diskmajmin) + + # diskdevpath has something like 253:0 + # and udev has put links in /dev/block/253:0 to the device name in /dev/ + return (diskdevpath, ptnum) + + +def devent2dev(devent): + if devent.startswith("/dev/"): + return devent + else: + result = util.get_mount_info(devent) + if not result: + raise ValueError("Could not determine device of '%s' % dev_ent") + return result[0] + + +def resize_devices(resizer, devices): + # returns a tuple of tuples containing (entry-in-devices, action, message) + info = [] + for devent in devices: + try: + blockdev = devent2dev(devent) + except ValueError as e: + info.append((devent, RESIZE.SKIPPED, + "unable to convert to device: %s" % e,)) + continue + + try: + statret = os.stat(blockdev) + except OSError as e: + info.append((devent, RESIZE.SKIPPED, + "stat of '%s' failed: %s" % (blockdev, e),)) + continue + + if not stat.S_ISBLK(statret.st_mode): + info.append((devent, RESIZE.SKIPPED, + "device '%s' not a block device" % blockdev,)) + continue + + try: + (disk, ptnum) = device_part_info(blockdev) + except (TypeError, ValueError) as e: + info.append((devent, RESIZE.SKIPPED, + "device_part_info(%s) failed: %s" % (blockdev, e),)) + continue + + try: + (old, new) = resizer.resize(disk, ptnum, blockdev) + if old == new: + info.append((devent, RESIZE.NOCHANGE, + "no change necessary (%s, %s)" % (disk, ptnum),)) + else: + info.append((devent, RESIZE.CHANGED, + "changed (%s, %s) from %s to %s" % + (disk, ptnum, old, new),)) + + except ResizeFailedException as e: + info.append((devent, RESIZE.FAILED, + "failed to resize: disk=%s, ptnum=%s: %s" % + (disk, ptnum, e),)) + + return info + + +def handle(_name, cfg, _cloud, log, _args): + if 'growpart' not in cfg: + log.debug("No 'growpart' entry in cfg. Using default: %s" % + DEFAULT_CONFIG) + cfg['growpart'] = DEFAULT_CONFIG + + mycfg = cfg.get('growpart') + if not isinstance(mycfg, dict): + log.warn("'growpart' in config was not a dict") + return + + mode = mycfg.get('mode', "auto") + if util.is_false(mode): + log.debug("growpart disabled: mode=%s" % mode) + return + + devices = util.get_cfg_option_list(cfg, "devices", ["/"]) + if not len(devices): + log.debug("growpart: empty device list") + return + + try: + resizer = resizer_factory(mode) + except (ValueError, TypeError) as e: + log.debug("growpart unable to find resizer for '%s': %s" % (mode, e)) + if mode != "auto": + raise e + return + + resized = util.log_time(logfunc=log.debug, msg="resize_devices", + func=resize_devices, args=(resizer, devices)) + for (entry, action, msg) in resized: + if action == RESIZE.CHANGED: + log.info("'%s' resized: %s" % (entry, msg)) + else: + log.debug("'%s' %s: %s" % (entry, action, msg)) + +# LP: 1212444 FIXME re-order and favor ResizeParted +#RESIZERS = (('growpart', ResizeGrowPart),) +RESIZERS = (('growpart', ResizeGrowPart), ('parted', ResizeParted)) diff --git a/cloudinit/config/cc_landscape.py b/cloudinit/config/cc_landscape.py index 2efdff79..8a709677 100644 --- a/cloudinit/config/cc_landscape.py +++ b/cloudinit/config/cc_landscape.py @@ -24,6 +24,7 @@ from StringIO import StringIO from configobj import ConfigObj +from cloudinit import type_utils from cloudinit import util from cloudinit.settings import PER_INSTANCE @@ -58,7 +59,8 @@ def handle(_name, cfg, cloud, log, _args): if not isinstance(ls_cloudcfg, (dict)): raise RuntimeError(("'landscape' key existed in config," " but not a dictionary type," - " is a %s instead"), util.obj_name(ls_cloudcfg)) + " is a %s instead"), + type_utils.obj_name(ls_cloudcfg)) if not ls_cloudcfg: return diff --git a/cloudinit/config/cc_mounts.py b/cloudinit/config/cc_mounts.py index 9010d97f..390ba711 100644 --- a/cloudinit/config/cc_mounts.py +++ b/cloudinit/config/cc_mounts.py @@ -22,6 +22,7 @@ from string import whitespace # pylint: disable=W0402 import re +from cloudinit import type_utils from cloudinit import util # Shortname matches 'sda', 'sda1', 'xvda', 'hda', 'sdb', xvdb, vda, vdd1, sr0 @@ -60,7 +61,7 @@ def handle(_name, cfg, cloud, log, _args): # skip something that wasn't a list if not isinstance(cfgmnt[i], list): log.warn("Mount option %s not a list, got a %s instead", - (i + 1), util.obj_name(cfgmnt[i])) + (i + 1), type_utils.obj_name(cfgmnt[i])) continue startname = str(cfgmnt[i][0]) diff --git a/cloudinit/config/cc_phone_home.py b/cloudinit/config/cc_phone_home.py index 886487f8..2e058ccd 100644 --- a/cloudinit/config/cc_phone_home.py +++ b/cloudinit/config/cc_phone_home.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2011 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # # Author: Scott Moser <scott.moser@canonical.com> # Author: Juerg Haefliger <juerg.haefliger@hp.com> @@ -19,7 +19,6 @@ # along with this program. If not, see <http://www.gnu.org/licenses/>. from cloudinit import templater -from cloudinit import url_helper as uhelp from cloudinit import util from cloudinit.settings import PER_INSTANCE @@ -66,8 +65,8 @@ def handle(name, cfg, cloud, log, args): tries = int(tries) except: tries = 10 - util.logexc(log, ("Configuration entry 'tries'" - " is not an integer, using %s instead"), tries) + util.logexc(log, "Configuration entry 'tries' is not an integer, " + "using %s instead", tries) if post_list == "all": post_list = POST_LIST_ALL @@ -86,8 +85,8 @@ def handle(name, cfg, cloud, log, args): try: all_keys[n] = util.load_file(path) except: - util.logexc(log, ("%s: failed to open, can not" - " phone home that data!"), path) + util.logexc(log, "%s: failed to open, can not phone home that " + "data!", path) submit_keys = {} for k in post_list: @@ -112,7 +111,9 @@ def handle(name, cfg, cloud, log, args): } url = templater.render_string(url, url_params) try: - uhelp.readurl(url, data=real_submit_keys, retries=tries, sec_between=3) + util.read_file_or_url(url, data=real_submit_keys, + retries=tries, sec_between=3, + ssl_details=util.fetch_ssl_details(cloud.paths)) except: - util.logexc(log, ("Failed to post phone home data to" - " %s in %s tries"), url, tries) + util.logexc(log, "Failed to post phone home data to %s in %s tries", + url, tries) diff --git a/cloudinit/config/cc_power_state_change.py b/cloudinit/config/cc_power_state_change.py index aefa3aff..188047e5 100644 --- a/cloudinit/config/cc_power_state_change.py +++ b/cloudinit/config/cc_power_state_change.py @@ -75,7 +75,7 @@ def load_power_state(cfg): ','.join(opt_map.keys())) delay = pstate.get("delay", "now") - if delay != "now" and not re.match("\+[0-9]+", delay): + if delay != "now" and not re.match(r"\+[0-9]+", delay): raise TypeError("power_state[delay] must be 'now' or '+m' (minutes).") args = ["shutdown", opt_map[mode], delay] @@ -100,7 +100,7 @@ def execmd(exe_args, output=None, data_in=None): proc = subprocess.Popen(exe_args, stdin=subprocess.PIPE, stdout=output, stderr=subprocess.STDOUT) proc.communicate(data_in) - ret = proc.returncode + ret = proc.returncode # pylint: disable=E1101 except Exception: doexit(EXIT_FAIL) doexit(ret) diff --git a/cloudinit/config/cc_resizefs.py b/cloudinit/config/cc_resizefs.py index 70294eda..56040fdd 100644 --- a/cloudinit/config/cc_resizefs.py +++ b/cloudinit/config/cc_resizefs.py @@ -18,50 +18,37 @@ # You should have received a copy of the GNU General Public License # along with this program. If not, see <http://www.gnu.org/licenses/>. +import errno import os import stat -import time from cloudinit.settings import PER_ALWAYS from cloudinit import util frequency = PER_ALWAYS -RESIZE_FS_PREFIXES_CMDS = [ - ('ext', 'resize2fs'), - ('xfs', 'xfs_growfs'), -] -NOBLOCK = "noblock" +def _resize_btrfs(mount_point, devpth): # pylint: disable=W0613 + return ('btrfs', 'filesystem', 'resize', 'max', mount_point) -def nodeify_path(devpth, where, log): - try: - st_dev = os.stat(where).st_dev - dev = os.makedev(os.major(st_dev), os.minor(st_dev)) - os.mknod(devpth, 0400 | stat.S_IFBLK, dev) - return st_dev - except: - if util.is_container(): - log.debug("Inside container, ignoring mknod failure in resizefs") - return - log.warn("Failed to make device node to resize %s at %s", - where, devpth) - raise +def _resize_ext(mount_point, devpth): # pylint: disable=W0613 + return ('resize2fs', devpth) -def get_fs_type(st_dev, path, log): - try: - dev_entries = util.find_devs_with(tag='TYPE', oformat='value', - no_cache=True, path=path) - if not dev_entries: - return None - return dev_entries[0].strip() - except util.ProcessExecutionError: - util.logexc(log, ("Failed to get filesystem type" - " of maj=%s, min=%s for path %s"), - os.major(st_dev), os.minor(st_dev), path) - raise +def _resize_xfs(mount_point, devpth): # pylint: disable=W0613 + return ('xfs_growfs', devpth) + +# Do not use a dictionary as these commands should be able to be used +# for multiple filesystem types if possible, e.g. one command for +# ext2, ext3 and ext4. +RESIZE_FS_PREFIXES_CMDS = [ + ('btrfs', _resize_btrfs), + ('ext', _resize_ext), + ('xfs', _resize_xfs), +] + +NOBLOCK = "noblock" def handle(name, cfg, _cloud, log, args): @@ -80,62 +67,77 @@ def handle(name, cfg, _cloud, log, args): # TODO(harlowja): allow what is to be resized to be configurable?? resize_what = "/" - with util.ExtendedTemporaryFile(prefix="cloudinit.resizefs.", - dir=resize_root_d, delete=True) as tfh: - devpth = tfh.name - - # Delete the file so that mknod will work - # but don't change the file handle to know that its - # removed so that when a later call that recreates - # occurs this temporary file will still benefit from - # auto deletion - tfh.unlink_now() - - st_dev = nodeify_path(devpth, resize_what, log) - fs_type = get_fs_type(st_dev, devpth, log) - if not fs_type: - log.warn("Could not determine filesystem type of %s", resize_what) - return - - resizer = None - fstype_lc = fs_type.lower() - for (pfix, root_cmd) in RESIZE_FS_PREFIXES_CMDS: - if fstype_lc.startswith(pfix): - resizer = root_cmd - break - - if not resizer: - log.warn("Not resizing unknown filesystem type %s for %s", - fs_type, resize_what) - return - - log.debug("Resizing %s (%s) using %s", resize_what, fs_type, resizer) - resize_cmd = [resizer, devpth] - - if resize_root == NOBLOCK: - # Fork to a child that will run - # the resize command - util.fork_cb(do_resize, resize_cmd, log) - # Don't delete the file now in the parent - tfh.delete = False + result = util.get_mount_info(resize_what, log) + if not result: + log.warn("Could not determine filesystem type of %s", resize_what) + return + + (devpth, fs_type, mount_point) = result + + # Ensure the path is a block device. + info = "dev=%s mnt_point=%s path=%s" % (devpth, mount_point, resize_what) + log.debug("resize_info: %s" % info) + + try: + statret = os.stat(devpth) + except OSError as exc: + if util.is_container() and exc.errno == errno.ENOENT: + log.debug("Device '%s' did not exist in container. " + "cannot resize: %s" % (devpth, info)) + elif exc.errno == errno.ENOENT: + log.warn("Device '%s' did not exist. cannot resize: %s" % + (devpth, info)) + else: + raise exc + return + + if not stat.S_ISBLK(statret.st_mode): + if util.is_container(): + log.debug("device '%s' not a block device in container." + " cannot resize: %s" % (devpth, info)) else: - do_resize(resize_cmd, log) + log.warn("device '%s' not a block device. cannot resize: %s" % + (devpth, info)) + return + + resizer = None + fstype_lc = fs_type.lower() + for (pfix, root_cmd) in RESIZE_FS_PREFIXES_CMDS: + if fstype_lc.startswith(pfix): + resizer = root_cmd + break + + if not resizer: + log.warn("Not resizing unknown filesystem type %s for %s", + fs_type, resize_what) + return + + resize_cmd = resizer(resize_what, devpth) + log.debug("Resizing %s (%s) using %s", resize_what, fs_type, + ' '.join(resize_cmd)) + + if resize_root == NOBLOCK: + # Fork to a child that will run + # the resize command + util.fork_cb( + util.log_time(logfunc=log.debug, msg="backgrounded Resizing", + func=do_resize, args=(resize_cmd, log))) + else: + util.log_time(logfunc=log.debug, msg="Resizing", + func=do_resize, args=(resize_cmd, log)) action = 'Resized' if resize_root == NOBLOCK: action = 'Resizing (via forking)' - log.debug("%s root filesystem (type=%s, maj=%i, min=%i, val=%s)", - action, fs_type, os.major(st_dev), os.minor(st_dev), resize_root) + log.debug("%s root filesystem (type=%s, val=%s)", action, fs_type, + resize_root) def do_resize(resize_cmd, log): - start = time.time() try: util.subp(resize_cmd) except util.ProcessExecutionError: util.logexc(log, "Failed to resize filesystem (cmd=%s)", resize_cmd) raise - tot_time = time.time() - start - log.debug("Resizing took %.3f seconds", tot_time) # TODO(harlowja): Should we add a fsck check after this to make # sure we didn't corrupt anything? diff --git a/cloudinit/config/cc_resolv_conf.py b/cloudinit/config/cc_resolv_conf.py index 8a460f7e..879b62b1 100644 --- a/cloudinit/config/cc_resolv_conf.py +++ b/cloudinit/config/cc_resolv_conf.py @@ -1,8 +1,10 @@ # vi: ts=4 expandtab # # Copyright (C) 2013 Craig Tracey +# Copyright (C) 2013 Hewlett-Packard Development Company, L.P. # # Author: Craig Tracey <craigtracey@gmail.com> +# Author: Juerg Haefliger <juerg.haefliger@hp.com> # # This program is free software: you can redistribute it and/or modify # it under the terms of the GNU General Public License version 3, as @@ -53,7 +55,7 @@ from cloudinit import util frequency = PER_INSTANCE -distros = ['fedora', 'rhel'] +distros = ['fedora', 'rhel', 'sles'] def generate_resolv_conf(cloud, log, params): diff --git a/cloudinit/config/cc_rightscale_userdata.py b/cloudinit/config/cc_rightscale_userdata.py index 4bf18516..c771728d 100644 --- a/cloudinit/config/cc_rightscale_userdata.py +++ b/cloudinit/config/cc_rightscale_userdata.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2011 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # # Author: Scott Moser <scott.moser@canonical.com> # Author: Juerg Haefliger <juerg.haefliger@hp.com> @@ -64,8 +64,8 @@ def handle(name, _cfg, cloud, log, _args): " raw userdata"), name, MY_HOOKNAME) return except: - util.logexc(log, ("Failed to parse query string %s" - " into a dictionary"), ud) + util.logexc(log, "Failed to parse query string %s into a dictionary", + ud) raise wrote_fns = [] @@ -86,8 +86,8 @@ def handle(name, _cfg, cloud, log, _args): wrote_fns.append(fname) except Exception as e: captured_excps.append(e) - util.logexc(log, "%s failed to read %s and write %s", - MY_NAME, url, fname) + util.logexc(log, "%s failed to read %s and write %s", MY_NAME, url, + fname) if wrote_fns: log.debug("Wrote out rightscale userdata to %s files", len(wrote_fns)) diff --git a/cloudinit/config/cc_set_hostname.py b/cloudinit/config/cc_set_hostname.py index 2b32fc94..5d7f4331 100644 --- a/cloudinit/config/cc_set_hostname.py +++ b/cloudinit/config/cc_set_hostname.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2011 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # # Author: Scott Moser <scott.moser@canonical.com> # Author: Juerg Haefliger <juerg.haefliger@hp.com> @@ -32,6 +32,6 @@ def handle(name, cfg, cloud, log, _args): log.debug("Setting the hostname to %s (%s)", fqdn, hostname) cloud.distro.set_hostname(hostname, fqdn) except Exception: - util.logexc(log, "Failed to set the hostname to %s (%s)", - fqdn, hostname) + util.logexc(log, "Failed to set the hostname to %s (%s)", fqdn, + hostname) raise diff --git a/cloudinit/config/cc_set_passwords.py b/cloudinit/config/cc_set_passwords.py index c6bf62fd..56a36906 100644 --- a/cloudinit/config/cc_set_passwords.py +++ b/cloudinit/config/cc_set_passwords.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2009-2010 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # # Author: Scott Moser <scott.moser@canonical.com> # Author: Juerg Haefliger <juerg.haefliger@hp.com> @@ -75,14 +75,14 @@ def handle(_name, cfg, cloud, log, args): plist_in.append("%s:%s" % (u, p)) users.append(u) - ch_in = '\n'.join(plist_in) + ch_in = '\n'.join(plist_in) + '\n' try: log.debug("Changing password for %s:", users) util.subp(['chpasswd'], ch_in) except Exception as e: errors.append(e) - util.logexc(log, - "Failed to set passwords with chpasswd for %s", users) + util.logexc(log, "Failed to set passwords with chpasswd for %s", + users) if len(randlist): blurb = ("Set the following 'random' passwords\n", diff --git a/cloudinit/config/cc_ssh.py b/cloudinit/config/cc_ssh.py index b623d476..64a5e3cb 100644 --- a/cloudinit/config/cc_ssh.py +++ b/cloudinit/config/cc_ssh.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2009-2010 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # # Author: Scott Moser <scott.moser@canonical.com> # Author: Juerg Haefliger <juerg.haefliger@hp.com> @@ -85,8 +85,8 @@ def handle(_name, cfg, cloud, log, _args): util.subp(cmd, capture=False) log.debug("Generated a key for %s from %s", pair[0], pair[1]) except: - util.logexc(log, ("Failed generated a key" - " for %s from %s"), pair[0], pair[1]) + util.logexc(log, "Failed generated a key for %s from %s", + pair[0], pair[1]) else: # if not, generate them genkeys = util.get_cfg_option_list(cfg, @@ -102,8 +102,8 @@ def handle(_name, cfg, cloud, log, _args): with util.SeLinuxGuard("/etc/ssh", recursive=True): util.subp(cmd, capture=False) except: - util.logexc(log, ("Failed generating key type" - " %s to file %s"), keytype, keyfile) + util.logexc(log, "Failed generating key type %s to " + "file %s", keytype, keyfile) try: (users, _groups) = ds.normalize_users_groups(cfg, cloud.distro) @@ -126,7 +126,7 @@ def apply_credentials(keys, user, disable_root, disable_root_opts): keys = set(keys) if user: - ssh_util.setup_user_keys(keys, user, '') + ssh_util.setup_user_keys(keys, user) if disable_root: if not user: @@ -135,4 +135,4 @@ def apply_credentials(keys, user, disable_root, disable_root_opts): else: key_prefix = '' - ssh_util.setup_user_keys(keys, 'root', key_prefix) + ssh_util.setup_user_keys(keys, 'root', options=key_prefix) diff --git a/cloudinit/config/cc_ssh_import_id.py b/cloudinit/config/cc_ssh_import_id.py index 83af36e9..50d96e15 100644 --- a/cloudinit/config/cc_ssh_import_id.py +++ b/cloudinit/config/cc_ssh_import_id.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2009-2010 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # # Author: Scott Moser <scott.moser@canonical.com> # Author: Juerg Haefliger <juerg.haefliger@hp.com> @@ -71,8 +71,8 @@ def handle(_name, cfg, cloud, log, args): try: import_ssh_ids(import_ids, user, log) except Exception as exc: - util.logexc(log, "ssh-import-id failed for: %s %s" % - (user, import_ids), exc) + util.logexc(log, "ssh-import-id failed for: %s %s", user, + import_ids) elist.append(exc) if len(elist): diff --git a/cloudinit/config/cc_update_hostname.py b/cloudinit/config/cc_update_hostname.py index 52225cd8..e396ba13 100644 --- a/cloudinit/config/cc_update_hostname.py +++ b/cloudinit/config/cc_update_hostname.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2011 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # # Author: Scott Moser <scott.moser@canonical.com> # Author: Juerg Haefliger <juerg.haefliger@hp.com> @@ -38,6 +38,6 @@ def handle(name, cfg, cloud, log, _args): log.debug("Updating hostname to %s (%s)", fqdn, hostname) cloud.distro.update_hostname(hostname, fqdn, prev_fn) except Exception: - util.logexc(log, "Failed to update the hostname to %s (%s)", - fqdn, hostname) + util.logexc(log, "Failed to update the hostname to %s (%s)", fqdn, + hostname) raise diff --git a/cloudinit/distros/__init__.py b/cloudinit/distros/__init__.py index 0db4aac7..74e95797 100644 --- a/cloudinit/distros/__init__.py +++ b/cloudinit/distros/__init__.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2012 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # Copyright (C) 2012 Yahoo! Inc. # # Author: Scott Moser <scott.moser@canonical.com> @@ -31,13 +31,15 @@ import re from cloudinit import importer from cloudinit import log as logging from cloudinit import ssh_util +from cloudinit import type_utils from cloudinit import util from cloudinit.distros.parsers import hosts OSFAMILIES = { 'debian': ['debian', 'ubuntu'], - 'redhat': ['fedora', 'rhel'] + 'redhat': ['fedora', 'rhel'], + 'suse': ['sles'] } LOG = logging.getLogger(__name__) @@ -45,9 +47,11 @@ LOG = logging.getLogger(__name__) class Distro(object): __metaclass__ = abc.ABCMeta + hosts_fn = "/etc/hosts" ci_sudoers_fn = "/etc/sudoers.d/90-cloud-init-users" hostname_conf_fn = "/etc/hostname" + tz_zone_dir = "/usr/share/zoneinfo" def __init__(self, name, cfg, paths): self._paths = paths @@ -64,6 +68,13 @@ class Distro(object): # to write this blob out in a distro format raise NotImplementedError() + def _find_tz_file(self, tz): + tz_file = os.path.join(self.tz_zone_dir, str(tz)) + if not os.path.isfile(tz_file): + raise IOError(("Invalid timezone %s," + " no file found at %s") % (tz, tz_file)) + return tz_file + def get_option(self, opt_name, default=None): return self._cfg.get(opt_name, default) @@ -73,7 +84,7 @@ class Distro(object): self._apply_hostname(hostname) @abc.abstractmethod - def package_command(self, cmd, args=None): + def package_command(self, cmd, args=None, pkgs=None): raise NotImplementedError() @abc.abstractmethod @@ -141,8 +152,8 @@ class Distro(object): try: util.subp(['hostname', hostname]) except util.ProcessExecutionError: - util.logexc(LOG, ("Failed to non-persistently adjust" - " the system hostname to %s"), hostname) + util.logexc(LOG, "Failed to non-persistently adjust the system " + "hostname to %s", hostname) @abc.abstractmethod def _select_hostname(self, hostname, fqdn): @@ -199,8 +210,8 @@ class Distro(object): try: self._write_hostname(hostname, fn) except IOError: - util.logexc(LOG, "Failed to write hostname %s to %s", - hostname, fn) + util.logexc(LOG, "Failed to write hostname %s to %s", hostname, + fn) if (sys_hostname and prev_hostname and sys_hostname != prev_hostname): @@ -280,15 +291,16 @@ class Distro(object): def get_default_user(self): return self.get_option('default_user') - def create_user(self, name, **kwargs): + def add_user(self, name, **kwargs): """ - Creates users for the system using the GNU passwd tools. This - will work on an GNU system. This should be overriden on - distros where useradd is not desirable or not available. + Add a user to the system using standard GNU tools """ + if util.is_user(name): + LOG.info("User %s already exists, skipping." % name) + return adduser_cmd = ['useradd', name] - x_adduser_cmd = ['useradd', name] + log_adduser_cmd = ['useradd', name] # Since we are creating users, we want to carefully validate the # inputs. If something goes wrong, we can end up with a system @@ -305,63 +317,65 @@ class Distro(object): "selinux_user": '--selinux-user', } - adduser_opts_flags = { + adduser_flags = { "no_user_group": '--no-user-group', "system": '--system', "no_log_init": '--no-log-init', - "no_create_home": "-M", } - redact_fields = ['passwd'] + redact_opts = ['passwd'] + + # Check the values and create the command + for key, val in kwargs.iteritems(): + + if key in adduser_opts and val and isinstance(val, str): + adduser_cmd.extend([adduser_opts[key], val]) - # Now check the value and create the command - for option in kwargs: - value = kwargs[option] - if option in adduser_opts and value \ - and isinstance(value, str): - adduser_cmd.extend([adduser_opts[option], value]) - # Redact certain fields from the logs - if option in redact_fields: - x_adduser_cmd.extend([adduser_opts[option], 'REDACTED']) - else: - x_adduser_cmd.extend([adduser_opts[option], value]) - elif option in adduser_opts_flags and value: - adduser_cmd.append(adduser_opts_flags[option]) # Redact certain fields from the logs - if option in redact_fields: - x_adduser_cmd.append('REDACTED') + if key in redact_opts: + log_adduser_cmd.extend([adduser_opts[key], 'REDACTED']) else: - x_adduser_cmd.append(adduser_opts_flags[option]) + log_adduser_cmd.extend([adduser_opts[key], val]) - # Default to creating home directory unless otherwise directed - # Also, we do not create home directories for system users. - if "no_create_home" not in kwargs and "system" not in kwargs: - adduser_cmd.append('-m') + elif key in adduser_flags and val: + adduser_cmd.append(adduser_flags[key]) + log_adduser_cmd.append(adduser_flags[key]) - # Create the user - if util.is_user(name): - LOG.warn("User %s already exists, skipping." % name) + # Don't create the home directory if directed so or if the user is a + # system user + if 'no_create_home' in kwargs or 'system' in kwargs: + adduser_cmd.append('-M') + log_adduser_cmd.append('-M') else: - LOG.debug("Adding user named %s", name) - try: - util.subp(adduser_cmd, logstring=x_adduser_cmd) - except Exception as e: - util.logexc(LOG, "Failed to create user %s due to error.", e) - raise e + adduser_cmd.append('-m') + log_adduser_cmd.append('-m') + + # Run the command + LOG.debug("Adding user %s", name) + try: + util.subp(adduser_cmd, logstring=log_adduser_cmd) + except Exception as e: + util.logexc(LOG, "Failed to create user %s", name) + raise e + + def create_user(self, name, **kwargs): + """ + Creates users for the system using the GNU passwd tools. This + will work on an GNU system. This should be overriden on + distros where useradd is not desirable or not available. + """ - # Set password if plain-text password provided + # Add the user + self.add_user(name, **kwargs) + + # Set password if plain-text password provided and non-empty if 'plain_text_passwd' in kwargs and kwargs['plain_text_passwd']: self.set_passwd(name, kwargs['plain_text_passwd']) # Default locking down the account. 'lock_passwd' defaults to True. # lock account unless lock_password is False. if kwargs.get('lock_passwd', True): - try: - util.subp(['passwd', '--lock', name]) - except Exception as e: - util.logexc(LOG, ("Failed to disable password logins for" - "user %s" % name), e) - raise e + self.lock_passwd(name) # Configure sudo access if 'sudo' in kwargs: @@ -370,21 +384,37 @@ class Distro(object): # Import SSH keys if 'ssh_authorized_keys' in kwargs: keys = set(kwargs['ssh_authorized_keys']) or [] - ssh_util.setup_user_keys(keys, name, key_prefix=None) + ssh_util.setup_user_keys(keys, name, options=None) return True + def lock_passwd(self, name): + """ + Lock the password of a user, i.e., disable password logins + """ + try: + # Need to use the short option name '-l' instead of '--lock' + # (which would be more descriptive) since SLES 11 doesn't know + # about long names. + util.subp(['passwd', '-l', name]) + except Exception as e: + util.logexc(LOG, 'Failed to disable password for user %s', name) + raise e + def set_passwd(self, user, passwd, hashed=False): pass_string = '%s:%s' % (user, passwd) cmd = ['chpasswd'] if hashed: - cmd.append('--encrypted') + # Need to use the short option name '-e' instead of '--encrypted' + # (which would be more descriptive) since SLES 11 doesn't know + # about long names. + cmd.append('-e') try: util.subp(cmd, pass_string, logstring="chpasswd for %s" % user) except Exception as e: - util.logexc(LOG, "Failed to set password for %s" % user) + util.logexc(LOG, "Failed to set password for %s", user) raise e return True @@ -426,7 +456,7 @@ class Distro(object): util.append_file(sudo_base, sudoers_contents) LOG.debug("Added '#includedir %s' to %s" % (path, sudo_base)) except IOError as e: - util.logexc(LOG, "Failed to write %s" % sudo_base, e) + util.logexc(LOG, "Failed to write %s", sudo_base) raise e util.ensure_dir(path, 0750) @@ -445,7 +475,7 @@ class Distro(object): lines.append("%s %s" % (user, rules)) else: msg = "Can not create sudoers rule addition with type %r" - raise TypeError(msg % (util.obj_name(rules))) + raise TypeError(msg % (type_utils.obj_name(rules))) content = "\n".join(lines) content += "\n" # trailing newline @@ -477,15 +507,15 @@ class Distro(object): try: util.subp(group_add_cmd) LOG.info("Created new group %s" % name) - except Exception as e: - util.logexc("Failed to create group %s" % name, e) + except Exception: + util.logexc("Failed to create group %s", name) # Add members to the group, if so defined if len(members) > 0: for member in members: if not util.is_user(member): LOG.warn("Unable to add group member '%s' to group '%s'" - "; user does not exist." % (member, name)) + "; user does not exist.", member, name) continue util.subp(['usermod', '-a', '-G', name, member]) @@ -568,7 +598,7 @@ def _normalize_groups(grp_cfg): c_grp_cfg[k] = [v] else: raise TypeError("Bad group member type %s" % - util.obj_name(v)) + type_utils.obj_name(v)) else: if isinstance(v, (list)): c_grp_cfg[k].extend(v) @@ -576,13 +606,13 @@ def _normalize_groups(grp_cfg): c_grp_cfg[k].append(v) else: raise TypeError("Bad group member type %s" % - util.obj_name(v)) + type_utils.obj_name(v)) elif isinstance(i, (str, basestring)): if i not in c_grp_cfg: c_grp_cfg[i] = [] else: raise TypeError("Unknown group name type %s" % - util.obj_name(i)) + type_utils.obj_name(i)) grp_cfg = c_grp_cfg groups = {} if isinstance(grp_cfg, (dict)): @@ -591,7 +621,7 @@ def _normalize_groups(grp_cfg): else: raise TypeError(("Group config must be list, dict " " or string types only and not %s") % - util.obj_name(grp_cfg)) + type_utils.obj_name(grp_cfg)) return groups @@ -622,7 +652,7 @@ def _normalize_users(u_cfg, def_user_cfg=None): ad_ucfg.append(v) else: raise TypeError(("Unmappable user value type %s" - " for key %s") % (util.obj_name(v), k)) + " for key %s") % (type_utils.obj_name(v), k)) u_cfg = ad_ucfg elif isinstance(u_cfg, (str, basestring)): u_cfg = util.uniq_merge_sorted(u_cfg) @@ -647,7 +677,7 @@ def _normalize_users(u_cfg, def_user_cfg=None): else: raise TypeError(("User config must be dictionary/list " " or string types only and not %s") % - util.obj_name(user_config)) + type_utils.obj_name(user_config)) # Ensure user options are in the right python friendly format if users: @@ -740,7 +770,7 @@ def normalize_users_groups(cfg, distro): } if not isinstance(old_user, (dict)): LOG.warn(("Format for 'user' key must be a string or " - "dictionary and not %s"), util.obj_name(old_user)) + "dictionary and not %s"), type_utils.obj_name(old_user)) old_user = {} # If no old user format, then assume the distro @@ -766,7 +796,7 @@ def normalize_users_groups(cfg, distro): if not isinstance(base_users, (list, dict, str, basestring)): LOG.warn(("Format for 'users' key must be a comma separated string" " or a dictionary or a list and not %s"), - util.obj_name(base_users)) + type_utils.obj_name(base_users)) base_users = [] if old_user: @@ -776,7 +806,7 @@ def normalize_users_groups(cfg, distro): # Just add it on at the end... base_users.append({'name': 'default'}) elif isinstance(base_users, (dict)): - base_users['default'] = base_users.get('default', True) + base_users['default'] = dict(base_users).get('default', True) elif isinstance(base_users, (str, basestring)): # Just append it on to be re-parsed later base_users += ",default" diff --git a/cloudinit/distros/debian.py b/cloudinit/distros/debian.py index 1a8e927b..8fe49cbe 100644 --- a/cloudinit/distros/debian.py +++ b/cloudinit/distros/debian.py @@ -33,6 +33,10 @@ from cloudinit.settings import PER_INSTANCE LOG = logging.getLogger(__name__) +APT_GET_COMMAND = ('apt-get', '--option=Dpkg::Options::=--force-confold', + '--option=Dpkg::options::=--force-unsafe-io', + '--assume-yes', '--quiet') + class Distro(distros.Distro): hostname_conf_fn = "/etc/hostname" @@ -40,7 +44,6 @@ class Distro(distros.Distro): network_conf_fn = "/etc/network/interfaces" tz_conf_fn = "/etc/timezone" tz_local_fn = "/etc/localtime" - tz_zone_dir = "/usr/share/zoneinfo" def __init__(self, name, cfg, paths): distros.Distro.__init__(self, name, cfg, paths) @@ -126,12 +129,7 @@ class Distro(distros.Distro): return "127.0.1.1" def set_timezone(self, tz): - # TODO(harlowja): move this code into - # the parent distro... - tz_file = os.path.join(self.tz_zone_dir, str(tz)) - if not os.path.isfile(tz_file): - raise RuntimeError(("Invalid timezone %s," - " no file found at %s") % (tz, tz_file)) + tz_file = self._find_tz_file(tz) # Note: "" provides trailing newline during join tz_lines = [ util.make_header(), @@ -142,20 +140,27 @@ class Distro(distros.Distro): # This ensures that the correct tz will be used for the system util.copy(tz_file, self.tz_local_fn) - def package_command(self, command, args=None, pkgs=[]): + def package_command(self, command, args=None, pkgs=None): + if pkgs is None: + pkgs = [] + e = os.environ.copy() # See: http://tiny.cc/kg91fw # Or: http://tiny.cc/mh91fw e['DEBIAN_FRONTEND'] = 'noninteractive' - cmd = ['apt-get', '--option', 'Dpkg::Options::=--force-confold', - '--assume-yes', '--quiet'] + cmd = list(self.get_option("apt_get_command", APT_GET_COMMAND)) if args and isinstance(args, str): cmd.append(args) elif args and isinstance(args, list): cmd.extend(args) - cmd.append(command) + subcmd = command + if command == "upgrade": + subcmd = self.get_option("apt_get_upgrade_subcommand", + "dist-upgrade") + + cmd.append(subcmd) pkglist = util.expand_package_list('%s=%s', pkgs) cmd.extend(pkglist) diff --git a/cloudinit/distros/parsers/resolv_conf.py b/cloudinit/distros/parsers/resolv_conf.py index 5733c25a..1be9d46b 100644 --- a/cloudinit/distros/parsers/resolv_conf.py +++ b/cloudinit/distros/parsers/resolv_conf.py @@ -137,8 +137,8 @@ class ResolvConf(object): self._contents.append(('option', ['search', s_list, ''])) return flat_sds - @local_domain.setter - def local_domain(self, domain): + @local_domain.setter # pl51222 pylint: disable=E1101 + def local_domain(self, domain): # pl51222 pylint: disable=E0102 self.parse() self._remove_option('domain') self._contents.append(('option', ['domain', str(domain), ''])) diff --git a/cloudinit/distros/rhel.py b/cloudinit/distros/rhel.py index 2f91e386..30195384 100644 --- a/cloudinit/distros/rhel.py +++ b/cloudinit/distros/rhel.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2012 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # Copyright (C) 2012 Yahoo! Inc. # # Author: Scott Moser <scott.moser@canonical.com> @@ -20,17 +20,12 @@ # You should have received a copy of the GNU General Public License # along with this program. If not, see <http://www.gnu.org/licenses/>. -import os - from cloudinit import distros - -from cloudinit.distros.parsers.resolv_conf import ResolvConf -from cloudinit.distros.parsers.sys_conf import SysConf - from cloudinit import helpers from cloudinit import log as logging from cloudinit import util +from cloudinit.distros import rhel_util from cloudinit.settings import PER_INSTANCE LOG = logging.getLogger(__name__) @@ -47,12 +42,13 @@ class Distro(distros.Distro): # See: http://tiny.cc/6r99fw clock_conf_fn = "/etc/sysconfig/clock" locale_conf_fn = '/etc/sysconfig/i18n' + systemd_locale_conf_fn = '/etc/locale.conf' network_conf_fn = "/etc/sysconfig/network" hostname_conf_fn = "/etc/sysconfig/network" + systemd_hostname_conf_fn = "/etc/hostname" network_script_tpl = '/etc/sysconfig/network-scripts/ifcfg-%s' resolve_conf_fn = "/etc/resolv.conf" tz_local_fn = "/etc/localtime" - tz_zone_dir = "/usr/share/zoneinfo" def __init__(self, name, cfg, paths): distros.Distro.__init__(self, name, cfg, paths) @@ -65,33 +61,9 @@ class Distro(distros.Distro): def install_packages(self, pkglist): self.package_command('install', pkgs=pkglist) - def _adjust_resolve(self, dns_servers, search_servers): - try: - r_conf = ResolvConf(util.load_file(self.resolve_conf_fn)) - r_conf.parse() - except IOError: - util.logexc(LOG, - "Failed at parsing %s reverting to an empty instance", - self.resolve_conf_fn) - r_conf = ResolvConf('') - r_conf.parse() - if dns_servers: - for s in dns_servers: - try: - r_conf.add_nameserver(s) - except ValueError: - util.logexc(LOG, "Failed at adding nameserver %s", s) - if search_servers: - for s in search_servers: - try: - r_conf.add_search_domain(s) - except ValueError: - util.logexc(LOG, "Failed at adding search domain %s", s) - util.write_file(self.resolve_conf_fn, str(r_conf), 0644) - def _write_network(self, settings): # TODO(harlowja) fix this... since this is the ubuntu format - entries = translate_network(settings) + entries = rhel_util.translate_network(settings) LOG.debug("Translated ubuntu style network settings %s into %s", settings, entries) # Make the intermediate format as the rhel format... @@ -110,54 +82,49 @@ class Distro(distros.Distro): 'MACADDR': info.get('hwaddress'), 'ONBOOT': _make_sysconfig_bool(info.get('auto')), } - self._update_sysconfig_file(net_fn, net_cfg) + rhel_util.update_sysconfig_file(net_fn, net_cfg) if 'dns-nameservers' in info: nameservers.extend(info['dns-nameservers']) if 'dns-search' in info: searchservers.extend(info['dns-search']) if nameservers or searchservers: - self._adjust_resolve(nameservers, searchservers) + rhel_util.update_resolve_conf_file(self.resolve_conf_fn, + nameservers, searchservers) if dev_names: net_cfg = { 'NETWORKING': _make_sysconfig_bool(True), } - self._update_sysconfig_file(self.network_conf_fn, net_cfg) + rhel_util.update_sysconfig_file(self.network_conf_fn, net_cfg) return dev_names - def _update_sysconfig_file(self, fn, adjustments, allow_empty=False): - if not adjustments: - return - (exists, contents) = self._read_conf(fn) - updated_am = 0 - for (k, v) in adjustments.items(): - if v is None: - continue - v = str(v) - if len(v) == 0 and not allow_empty: - continue - contents[k] = v - updated_am += 1 - if updated_am: - lines = [ - str(contents), - ] - if not exists: - lines.insert(0, util.make_header()) - util.write_file(fn, "\n".join(lines), 0644) + def _dist_uses_systemd(self): + # Fedora 18 and RHEL 7 were the first adopters in their series + (dist, vers) = util.system_info()['dist'][:2] + major = (int)(vers.split('.')[0]) + return ((dist.startswith('Red Hat Enterprise Linux') and major >= 7) + or (dist.startswith('Fedora') and major >= 18)) def apply_locale(self, locale, out_fn=None): - if not out_fn: - out_fn = self.locale_conf_fn + if self._dist_uses_systemd(): + if not out_fn: + out_fn = self.systemd_locale_conf_fn + out_fn = self.systemd_locale_conf_fn + else: + if not out_fn: + out_fn = self.locale_conf_fn locale_cfg = { 'LANG': locale, } - self._update_sysconfig_file(out_fn, locale_cfg) + rhel_util.update_sysconfig_file(out_fn, locale_cfg) def _write_hostname(self, hostname, out_fn): - host_cfg = { - 'HOSTNAME': hostname, - } - self._update_sysconfig_file(out_fn, host_cfg) + if self._dist_uses_systemd(): + util.subp(['hostnamectl', 'set-hostname', str(hostname)]) + else: + host_cfg = { + 'HOSTNAME': hostname, + } + rhel_util.update_sysconfig_file(out_fn, host_cfg) def _select_hostname(self, hostname, fqdn): # See: http://bit.ly/TwitgL @@ -167,25 +134,25 @@ class Distro(distros.Distro): return hostname def _read_system_hostname(self): - return (self.network_conf_fn, - self._read_hostname(self.network_conf_fn)) + if self._dist_uses_systemd(): + host_fn = self.systemd_hostname_conf_fn + else: + host_fn = self.hostname_conf_fn + return (host_fn, self._read_hostname(host_fn)) def _read_hostname(self, filename, default=None): - (_exists, contents) = self._read_conf(filename) - if 'HOSTNAME' in contents: - return contents['HOSTNAME'] + if self._dist_uses_systemd(): + (out, _err) = util.subp(['hostname']) + if len(out): + return out + else: + return default else: - return default - - def _read_conf(self, fn): - exists = False - try: - contents = util.load_file(fn).splitlines() - exists = True - except IOError: - contents = [] - return (exists, - SysConf(contents)) + (_exists, contents) = rhel_util.read_sysconfig_file(filename) + if 'HOSTNAME' in contents: + return contents['HOSTNAME'] + else: + return default def _bring_up_interfaces(self, device_names): if device_names and 'all' in device_names: @@ -194,21 +161,25 @@ class Distro(distros.Distro): return distros.Distro._bring_up_interfaces(self, device_names) def set_timezone(self, tz): - # TODO(harlowja): move this code into - # the parent distro... - tz_file = os.path.join(self.tz_zone_dir, str(tz)) - if not os.path.isfile(tz_file): - raise RuntimeError(("Invalid timezone %s," - " no file found at %s") % (tz, tz_file)) - # Adjust the sysconfig clock zone setting - clock_cfg = { - 'ZONE': str(tz), - } - self._update_sysconfig_file(self.clock_conf_fn, clock_cfg) - # This ensures that the correct tz will be used for the system - util.copy(tz_file, self.tz_local_fn) + tz_file = self._find_tz_file(tz) + if self._dist_uses_systemd(): + # Currently, timedatectl complains if invoked during startup + # so for compatibility, create the link manually. + util.del_file(self.tz_local_fn) + util.sym_link(tz_file, self.tz_local_fn) + else: + # Adjust the sysconfig clock zone setting + clock_cfg = { + 'ZONE': str(tz), + } + rhel_util.update_sysconfig_file(self.clock_conf_fn, clock_cfg) + # This ensures that the correct tz will be used for the system + util.copy(tz_file, self.tz_local_fn) + + def package_command(self, command, args=None, pkgs=None): + if pkgs is None: + pkgs = [] - def package_command(self, command, args=None, pkgs=[]): cmd = ['yum'] # If enabled, then yum will be tolerant of errors on the command line # with regard to packages. @@ -236,90 +207,3 @@ class Distro(distros.Distro): def update_package_sources(self): self._runner.run("update-sources", self.package_command, ["makecache"], freq=PER_INSTANCE) - - -# This is a util function to translate a ubuntu /etc/network/interfaces 'blob' -# to a rhel equiv. that can then be written to /etc/sysconfig/network-scripts/ -# TODO(harlowja) remove when we have python-netcf active... -def translate_network(settings): - # Get the standard cmd, args from the ubuntu format - entries = [] - for line in settings.splitlines(): - line = line.strip() - if not line or line.startswith("#"): - continue - split_up = line.split(None, 1) - if len(split_up) <= 1: - continue - entries.append(split_up) - # Figure out where each iface section is - ifaces = [] - consume = {} - for (cmd, args) in entries: - if cmd == 'iface': - if consume: - ifaces.append(consume) - consume = {} - consume[cmd] = args - else: - consume[cmd] = args - # Check if anything left over to consume - absorb = False - for (cmd, args) in consume.iteritems(): - if cmd == 'iface': - absorb = True - if absorb: - ifaces.append(consume) - # Now translate - real_ifaces = {} - for info in ifaces: - if 'iface' not in info: - continue - iface_details = info['iface'].split(None) - dev_name = None - if len(iface_details) >= 1: - dev = iface_details[0].strip().lower() - if dev: - dev_name = dev - if not dev_name: - continue - iface_info = {} - if len(iface_details) >= 3: - proto_type = iface_details[2].strip().lower() - # Seems like this can be 'loopback' which we don't - # really care about - if proto_type in ['dhcp', 'static']: - iface_info['bootproto'] = proto_type - # These can just be copied over - for k in ['netmask', 'address', 'gateway', 'broadcast']: - if k in info: - val = info[k].strip().lower() - if val: - iface_info[k] = val - # Name server info provided?? - if 'dns-nameservers' in info: - iface_info['dns-nameservers'] = info['dns-nameservers'].split() - # Name server search info provided?? - if 'dns-search' in info: - iface_info['dns-search'] = info['dns-search'].split() - # Is any mac address spoofing going on?? - if 'hwaddress' in info: - hw_info = info['hwaddress'].lower().strip() - hw_split = hw_info.split(None, 1) - if len(hw_split) == 2 and hw_split[0].startswith('ether'): - hw_addr = hw_split[1] - if hw_addr: - iface_info['hwaddress'] = hw_addr - real_ifaces[dev_name] = iface_info - # Check for those that should be started on boot via 'auto' - for (cmd, args) in entries: - if cmd == 'auto': - # Seems like auto can be like 'auto eth0 eth0:1' so just get the - # first part out as the device name - args = args.split(None) - if not args: - continue - dev_name = args[0].strip().lower() - if dev_name in real_ifaces: - real_ifaces[dev_name]['auto'] = True - return real_ifaces diff --git a/cloudinit/distros/rhel_util.py b/cloudinit/distros/rhel_util.py new file mode 100644 index 00000000..1aba58b8 --- /dev/null +++ b/cloudinit/distros/rhel_util.py @@ -0,0 +1,177 @@ +# vi: ts=4 expandtab +# +# Copyright (C) 2012 Canonical Ltd. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012 Yahoo! Inc. +# +# Author: Scott Moser <scott.moser@canonical.com> +# Author: Juerg Haefliger <juerg.haefliger@hp.com> +# Author: Joshua Harlow <harlowja@yahoo-inc.com> +# +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License version 3, as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. +# + +from cloudinit.distros.parsers.resolv_conf import ResolvConf +from cloudinit.distros.parsers.sys_conf import SysConf + +from cloudinit import log as logging +from cloudinit import util + +LOG = logging.getLogger(__name__) + + +# This is a util function to translate Debian based distro interface blobs as +# given in /etc/network/interfaces to an equivalent format for distributions +# that use ifcfg-* style (Red Hat and SUSE). +# TODO(harlowja) remove when we have python-netcf active... +def translate_network(settings): + # Get the standard cmd, args from the ubuntu format + entries = [] + for line in settings.splitlines(): + line = line.strip() + if not line or line.startswith("#"): + continue + split_up = line.split(None, 1) + if len(split_up) <= 1: + continue + entries.append(split_up) + # Figure out where each iface section is + ifaces = [] + consume = {} + for (cmd, args) in entries: + if cmd == 'iface': + if consume: + ifaces.append(consume) + consume = {} + consume[cmd] = args + else: + consume[cmd] = args + # Check if anything left over to consume + absorb = False + for (cmd, args) in consume.iteritems(): + if cmd == 'iface': + absorb = True + if absorb: + ifaces.append(consume) + # Now translate + real_ifaces = {} + for info in ifaces: + if 'iface' not in info: + continue + iface_details = info['iface'].split(None) + dev_name = None + if len(iface_details) >= 1: + dev = iface_details[0].strip().lower() + if dev: + dev_name = dev + if not dev_name: + continue + iface_info = {} + if len(iface_details) >= 3: + proto_type = iface_details[2].strip().lower() + # Seems like this can be 'loopback' which we don't + # really care about + if proto_type in ['dhcp', 'static']: + iface_info['bootproto'] = proto_type + # These can just be copied over + for k in ['netmask', 'address', 'gateway', 'broadcast']: + if k in info: + val = info[k].strip().lower() + if val: + iface_info[k] = val + # Name server info provided?? + if 'dns-nameservers' in info: + iface_info['dns-nameservers'] = info['dns-nameservers'].split() + # Name server search info provided?? + if 'dns-search' in info: + iface_info['dns-search'] = info['dns-search'].split() + # Is any mac address spoofing going on?? + if 'hwaddress' in info: + hw_info = info['hwaddress'].lower().strip() + hw_split = hw_info.split(None, 1) + if len(hw_split) == 2 and hw_split[0].startswith('ether'): + hw_addr = hw_split[1] + if hw_addr: + iface_info['hwaddress'] = hw_addr + real_ifaces[dev_name] = iface_info + # Check for those that should be started on boot via 'auto' + for (cmd, args) in entries: + if cmd == 'auto': + # Seems like auto can be like 'auto eth0 eth0:1' so just get the + # first part out as the device name + args = args.split(None) + if not args: + continue + dev_name = args[0].strip().lower() + if dev_name in real_ifaces: + real_ifaces[dev_name]['auto'] = True + return real_ifaces + + +# Helper function to update a RHEL/SUSE /etc/sysconfig/* file +def update_sysconfig_file(fn, adjustments, allow_empty=False): + if not adjustments: + return + (exists, contents) = read_sysconfig_file(fn) + updated_am = 0 + for (k, v) in adjustments.items(): + if v is None: + continue + v = str(v) + if len(v) == 0 and not allow_empty: + continue + contents[k] = v + updated_am += 1 + if updated_am: + lines = [ + str(contents), + ] + if not exists: + lines.insert(0, util.make_header()) + util.write_file(fn, "\n".join(lines) + "\n", 0644) + + +# Helper function to read a RHEL/SUSE /etc/sysconfig/* file +def read_sysconfig_file(fn): + exists = False + try: + contents = util.load_file(fn).splitlines() + exists = True + except IOError: + contents = [] + return (exists, SysConf(contents)) + + +# Helper function to update RHEL/SUSE /etc/resolv.conf +def update_resolve_conf_file(fn, dns_servers, search_servers): + try: + r_conf = ResolvConf(util.load_file(fn)) + r_conf.parse() + except IOError: + util.logexc(LOG, "Failed at parsing %s reverting to an empty " + "instance", fn) + r_conf = ResolvConf('') + r_conf.parse() + if dns_servers: + for s in dns_servers: + try: + r_conf.add_nameserver(s) + except ValueError: + util.logexc(LOG, "Failed at adding nameserver %s", s) + if search_servers: + for s in search_servers: + try: + r_conf.add_search_domain(s) + except ValueError: + util.logexc(LOG, "Failed at adding search domain %s", s) + util.write_file(fn, str(r_conf), 0644) diff --git a/cloudinit/distros/sles.py b/cloudinit/distros/sles.py new file mode 100644 index 00000000..f2ac4efc --- /dev/null +++ b/cloudinit/distros/sles.py @@ -0,0 +1,185 @@ +# vi: ts=4 expandtab +# +# Copyright (C) 2013 Hewlett-Packard Development Company, L.P. +# +# Author: Juerg Haefliger <juerg.haefliger@hp.com> +# +# Leaning very heavily on the RHEL and Debian implementation +# +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License version 3, as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. + +from cloudinit import distros + +from cloudinit.distros.parsers.hostname import HostnameConf + +from cloudinit import helpers +from cloudinit import log as logging +from cloudinit import util + +from cloudinit.distros import rhel_util +from cloudinit.settings import PER_INSTANCE + +LOG = logging.getLogger(__name__) + + +class Distro(distros.Distro): + clock_conf_fn = '/etc/sysconfig/clock' + locale_conf_fn = '/etc/sysconfig/language' + network_conf_fn = '/etc/sysconfig/network' + hostname_conf_fn = '/etc/HOSTNAME' + network_script_tpl = '/etc/sysconfig/network/ifcfg-%s' + resolve_conf_fn = '/etc/resolv.conf' + tz_local_fn = '/etc/localtime' + + def __init__(self, name, cfg, paths): + distros.Distro.__init__(self, name, cfg, paths) + # This will be used to restrict certain + # calls from repeatly happening (when they + # should only happen say once per instance...) + self._runner = helpers.Runners(paths) + self.osfamily = 'suse' + + def install_packages(self, pkglist): + self.package_command('install', args='-l', pkgs=pkglist) + + def _write_network(self, settings): + # Convert debian settings to ifcfg format + entries = rhel_util.translate_network(settings) + LOG.debug("Translated ubuntu style network settings %s into %s", + settings, entries) + # Make the intermediate format as the suse format... + nameservers = [] + searchservers = [] + dev_names = entries.keys() + for (dev, info) in entries.iteritems(): + net_fn = self.network_script_tpl % (dev) + mode = info.get('auto') + if mode and mode.lower() == 'true': + mode = 'auto' + else: + mode = 'manual' + net_cfg = { + 'BOOTPROTO': info.get('bootproto'), + 'BROADCAST': info.get('broadcast'), + 'GATEWAY': info.get('gateway'), + 'IPADDR': info.get('address'), + 'LLADDR': info.get('hwaddress'), + 'NETMASK': info.get('netmask'), + 'STARTMODE': mode, + 'USERCONTROL': 'no' + } + if dev != 'lo': + net_cfg['ETHERDEVICE'] = dev + net_cfg['ETHTOOL_OPTIONS'] = '' + else: + net_cfg['FIREWALL'] = 'no' + rhel_util.update_sysconfig_file(net_fn, net_cfg, True) + if 'dns-nameservers' in info: + nameservers.extend(info['dns-nameservers']) + if 'dns-search' in info: + searchservers.extend(info['dns-search']) + if nameservers or searchservers: + rhel_util.update_resolve_conf_file(self.resolve_conf_fn, + nameservers, searchservers) + return dev_names + + def apply_locale(self, locale, out_fn=None): + if not out_fn: + out_fn = self.locale_conf_fn + locale_cfg = { + 'RC_LANG': locale, + } + rhel_util.update_sysconfig_file(out_fn, locale_cfg) + + def _write_hostname(self, hostname, out_fn): + conf = None + try: + # Try to update the previous one + # so lets see if we can read it first. + conf = self._read_hostname_conf(out_fn) + except IOError: + pass + if not conf: + conf = HostnameConf('') + conf.set_hostname(hostname) + util.write_file(out_fn, str(conf), 0644) + + def _select_hostname(self, hostname, fqdn): + # Prefer the short hostname over the long + # fully qualified domain name + if not hostname: + return fqdn + return hostname + + def _read_system_hostname(self): + host_fn = self.hostname_conf_fn + return (host_fn, self._read_hostname(host_fn)) + + def _read_hostname_conf(self, filename): + conf = HostnameConf(util.load_file(filename)) + conf.parse() + return conf + + def _read_hostname(self, filename, default=None): + hostname = None + try: + conf = self._read_hostname_conf(filename) + hostname = conf.hostname + except IOError: + pass + if not hostname: + return default + return hostname + + def _bring_up_interfaces(self, device_names): + if device_names and 'all' in device_names: + raise RuntimeError(('Distro %s can not translate ' + 'the device name "all"') % (self.name)) + return distros.Distro._bring_up_interfaces(self, device_names) + + def set_timezone(self, tz): + tz_file = self._find_tz_file(tz) + # Adjust the sysconfig clock zone setting + clock_cfg = { + 'TIMEZONE': str(tz), + } + rhel_util.update_sysconfig_file(self.clock_conf_fn, clock_cfg) + # This ensures that the correct tz will be used for the system + util.copy(tz_file, self.tz_local_fn) + + def package_command(self, command, args=None, pkgs=None): + if pkgs is None: + pkgs = [] + + cmd = ['zypper'] + # No user interaction possible, enable non-interactive mode + cmd.append('--non-interactive') + + # Comand is the operation, such as install + cmd.append(command) + + # args are the arguments to the command, not global options + if args and isinstance(args, str): + cmd.append(args) + elif args and isinstance(args, list): + cmd.extend(args) + + pkglist = util.expand_package_list('%s-%s', pkgs) + cmd.extend(pkglist) + + # Allow the output of this to flow outwards (ie not be captured) + util.subp(cmd, capture=False) + + def update_package_sources(self): + self._runner.run("update-sources", self.package_command, + ['refresh'], freq=PER_INSTANCE) diff --git a/cloudinit/ec2_utils.py b/cloudinit/ec2_utils.py index 46b93f39..fcd511c5 100644 --- a/cloudinit/ec2_utils.py +++ b/cloudinit/ec2_utils.py @@ -28,6 +28,10 @@ import boto.utils as boto_utils # would have existed) do not exist due to the blocking # that occurred. +# TODO(harlowja): https://github.com/boto/boto/issues/1401 +# When boto finally moves to using requests, we should be able +# to provide it ssl details, it does not yet, so we can't provide them... + def _unlazy_dict(mp): if not isinstance(mp, (dict)): diff --git a/cloudinit/handlers/__init__.py b/cloudinit/handlers/__init__.py index 8d6dcd4d..2ddc75f4 100644 --- a/cloudinit/handlers/__init__.py +++ b/cloudinit/handlers/__init__.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2012 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # Copyright (C) 2012 Yahoo! Inc. # # Author: Scott Moser <scott.moser@canonical.com> @@ -27,6 +27,7 @@ from cloudinit.settings import (PER_ALWAYS, PER_INSTANCE, FREQUENCIES) from cloudinit import importer from cloudinit import log as logging +from cloudinit import type_utils from cloudinit import util LOG = logging.getLogger(__name__) @@ -61,6 +62,7 @@ INCLUSION_TYPES_MAP = { '#part-handler': 'text/part-handler', '#cloud-boothook': 'text/cloud-boothook', '#cloud-config-archive': 'text/cloud-config-archive', + '#cloud-config-jsonp': 'text/cloud-config-jsonp', } # Sorted longest first @@ -69,7 +71,6 @@ INCLUSION_SRCH = sorted(list(INCLUSION_TYPES_MAP.keys()), class Handler(object): - __metaclass__ = abc.ABCMeta def __init__(self, frequency, version=2): @@ -77,53 +78,65 @@ class Handler(object): self.frequency = frequency def __repr__(self): - return "%s: [%s]" % (util.obj_name(self), self.list_types()) + return "%s: [%s]" % (type_utils.obj_name(self), self.list_types()) @abc.abstractmethod def list_types(self): raise NotImplementedError() - def handle_part(self, data, ctype, filename, payload, frequency): - return self._handle_part(data, ctype, filename, payload, frequency) - @abc.abstractmethod - def _handle_part(self, data, ctype, filename, payload, frequency): + def handle_part(self, *args, **kwargs): raise NotImplementedError() -def run_part(mod, data, ctype, filename, payload, frequency): +def run_part(mod, data, filename, payload, frequency, headers): mod_freq = mod.frequency if not (mod_freq == PER_ALWAYS or (frequency == PER_INSTANCE and mod_freq == PER_INSTANCE)): return - mod_ver = mod.handler_version # Sanity checks on version (should be an int convertable) try: + mod_ver = mod.handler_version mod_ver = int(mod_ver) - except: + except (TypeError, ValueError, AttributeError): mod_ver = 1 + content_type = headers['Content-Type'] try: LOG.debug("Calling handler %s (%s, %s, %s) with frequency %s", - mod, ctype, filename, mod_ver, frequency) - if mod_ver >= 2: + mod, content_type, filename, mod_ver, frequency) + if mod_ver == 3: + # Treat as v. 3 which does get a frequency + headers + mod.handle_part(data, content_type, filename, + payload, frequency, headers) + elif mod_ver == 2: # Treat as v. 2 which does get a frequency - mod.handle_part(data, ctype, filename, payload, frequency) - else: + mod.handle_part(data, content_type, filename, + payload, frequency) + elif mod_ver == 1: # Treat as v. 1 which gets no frequency - mod.handle_part(data, ctype, filename, payload) + mod.handle_part(data, content_type, filename, payload) + else: + raise ValueError("Unknown module version %s" % (mod_ver)) except: - util.logexc(LOG, ("Failed calling handler %s (%s, %s, %s)" - " with frequency %s"), - mod, ctype, filename, - mod_ver, frequency) + util.logexc(LOG, "Failed calling handler %s (%s, %s, %s) with " + "frequency %s", mod, content_type, filename, mod_ver, + frequency) def call_begin(mod, data, frequency): - run_part(mod, data, CONTENT_START, None, None, frequency) + # Create a fake header set + headers = { + 'Content-Type': CONTENT_START, + } + run_part(mod, data, None, None, frequency, headers) def call_end(mod, data, frequency): - run_part(mod, data, CONTENT_END, None, None, frequency) + # Create a fake header set + headers = { + 'Content-Type': CONTENT_END, + } + run_part(mod, data, None, None, frequency, headers) def walker_handle_handler(pdata, _ctype, _filename, payload): @@ -139,14 +152,13 @@ def walker_handle_handler(pdata, _ctype, _filename, payload): try: mod = fixup_handler(importer.import_module(modname)) call_begin(mod, pdata['data'], frequency) - # Only register and increment - # after the above have worked (so we don't if it - # fails) - handlers.register(mod) + # Only register and increment after the above have worked, so we don't + # register if it fails starting. + handlers.register(mod, initialized=True) pdata['handlercount'] = curcount + 1 except: - util.logexc(LOG, ("Failed at registering python file: %s" - " (part handler %s)"), modfname, curcount) + util.logexc(LOG, "Failed at registering python file: %s (part " + "handler %s)", modfname, curcount) def _extract_first_or_bytes(blob, size): @@ -173,26 +185,27 @@ def _escape_string(text): return text -def walker_callback(pdata, ctype, filename, payload): - if ctype in PART_CONTENT_TYPES: - walker_handle_handler(pdata, ctype, filename, payload) +def walker_callback(data, filename, payload, headers): + content_type = headers['Content-Type'] + if content_type in PART_CONTENT_TYPES: + walker_handle_handler(data, content_type, filename, payload) return - handlers = pdata['handlers'] - if ctype in pdata['handlers']: - run_part(handlers[ctype], pdata['data'], ctype, filename, - payload, pdata['frequency']) + handlers = data['handlers'] + if content_type in handlers: + run_part(handlers[content_type], data['data'], filename, + payload, data['frequency'], headers) elif payload: # Extract the first line or 24 bytes for displaying in the log start = _extract_first_or_bytes(payload, 24) details = "'%s...'" % (_escape_string(start)) - if ctype == NOT_MULTIPART_TYPE: + if content_type == NOT_MULTIPART_TYPE: LOG.warning("Unhandled non-multipart (%s) userdata: %s", - ctype, details) + content_type, details) else: LOG.warning("Unhandled unknown content-type (%s) userdata: %s", - ctype, details) + content_type, details) else: - LOG.debug("empty payload of type %s" % ctype) + LOG.debug("Empty payload of type %s", content_type) # Callback is a function that will be called with @@ -212,7 +225,10 @@ def walk(msg, callback, data): if not filename: filename = PART_FN_TPL % (partnum) - callback(data, ctype, filename, part.get_payload(decode=True)) + headers = dict(part) + LOG.debug(headers) + headers['Content-Type'] = ctype + callback(data, filename, part.get_payload(decode=True), headers) partnum = partnum + 1 diff --git a/cloudinit/handlers/boot_hook.py b/cloudinit/handlers/boot_hook.py index 456b8020..1848ce2c 100644 --- a/cloudinit/handlers/boot_hook.py +++ b/cloudinit/handlers/boot_hook.py @@ -29,6 +29,7 @@ from cloudinit import util from cloudinit.settings import (PER_ALWAYS) LOG = logging.getLogger(__name__) +BOOTHOOK_PREFIX = "#cloud-boothook" class BootHookPartHandler(handlers.Handler): @@ -41,22 +42,19 @@ class BootHookPartHandler(handlers.Handler): def list_types(self): return [ - handlers.type_from_starts_with("#cloud-boothook"), + handlers.type_from_starts_with(BOOTHOOK_PREFIX), ] def _write_part(self, payload, filename): filename = util.clean_filename(filename) - payload = util.dos2unix(payload) - prefix = "#cloud-boothook" - start = 0 - if payload.startswith(prefix): - start = len(prefix) + 1 filepath = os.path.join(self.boothook_dir, filename) - contents = payload[start:] - util.write_file(filepath, contents, 0700) + contents = util.strip_prefix_suffix(util.dos2unix(payload), + prefix=BOOTHOOK_PREFIX) + util.write_file(filepath, contents.lstrip(), 0700) return filepath - def _handle_part(self, _data, ctype, filename, payload, _frequency): + def handle_part(self, _data, ctype, filename, # pylint: disable=W0221 + payload, frequency): # pylint: disable=W0613 if ctype in handlers.CONTENT_SIGNALS: return @@ -69,5 +67,5 @@ class BootHookPartHandler(handlers.Handler): except util.ProcessExecutionError: util.logexc(LOG, "Boothooks script %s execution error", filepath) except Exception: - util.logexc(LOG, ("Boothooks unknown " - "error when running %s"), filepath) + util.logexc(LOG, "Boothooks unknown error when running %s", + filepath) diff --git a/cloudinit/handlers/cloud_config.py b/cloudinit/handlers/cloud_config.py index f6d95244..34a73115 100644 --- a/cloudinit/handlers/cloud_config.py +++ b/cloudinit/handlers/cloud_config.py @@ -20,43 +20,143 @@ # You should have received a copy of the GNU General Public License # along with this program. If not, see <http://www.gnu.org/licenses/>. +import jsonpatch + from cloudinit import handlers from cloudinit import log as logging +from cloudinit import mergers from cloudinit import util from cloudinit.settings import (PER_ALWAYS) LOG = logging.getLogger(__name__) +MERGE_HEADER = 'Merge-Type' + +# Due to the way the loading of yaml configuration was done previously, +# where previously each cloud config part was appended to a larger yaml +# file and then finally that file was loaded as one big yaml file we need +# to mimic that behavior by altering the default strategy to be replacing +# keys of prior merges. +# +# +# For example +# #file 1 +# a: 3 +# #file 2 +# a: 22 +# #combined file (comments not included) +# a: 3 +# a: 22 +# +# This gets loaded into yaml with final result {'a': 22} +DEF_MERGERS = mergers.string_extract_mergers('dict(replace)+list()+str()') +CLOUD_PREFIX = "#cloud-config" +JSONP_PREFIX = "#cloud-config-jsonp" + +# The file header -> content types this module will handle. +CC_TYPES = { + JSONP_PREFIX: handlers.type_from_starts_with(JSONP_PREFIX), + CLOUD_PREFIX: handlers.type_from_starts_with(CLOUD_PREFIX), +} + class CloudConfigPartHandler(handlers.Handler): def __init__(self, paths, **_kwargs): - handlers.Handler.__init__(self, PER_ALWAYS) - self.cloud_buf = [] + handlers.Handler.__init__(self, PER_ALWAYS, version=3) + self.cloud_buf = None self.cloud_fn = paths.get_ipath("cloud_config") + self.file_names = [] def list_types(self): - return [ - handlers.type_from_starts_with("#cloud-config"), - ] + return list(CC_TYPES.values()) - def _write_cloud_config(self, buf): + def _write_cloud_config(self): if not self.cloud_fn: return - lines = [str(b) for b in buf] - payload = "\n".join(lines) - util.write_file(self.cloud_fn, payload, 0600) + # Capture which files we merged from... + file_lines = [] + if self.file_names: + file_lines.append("# from %s files" % (len(self.file_names))) + for fn in self.file_names: + if not fn: + fn = '?' + file_lines.append("# %s" % (fn)) + file_lines.append("") + if self.cloud_buf is not None: + # Something was actually gathered.... + lines = [ + CLOUD_PREFIX, + '', + ] + lines.extend(file_lines) + lines.append(util.yaml_dumps(self.cloud_buf)) + else: + lines = [] + util.write_file(self.cloud_fn, "\n".join(lines), 0600) + + def _extract_mergers(self, payload, headers): + merge_header_headers = '' + for h in [MERGE_HEADER, 'X-%s' % (MERGE_HEADER)]: + tmp_h = headers.get(h, '') + if tmp_h: + merge_header_headers = tmp_h + break + # Select either the merge-type from the content + # or the merge type from the headers or default to our own set + # if neither exists (or is empty) from the later. + payload_yaml = util.load_yaml(payload) + mergers_yaml = mergers.dict_extract_mergers(payload_yaml) + mergers_header = mergers.string_extract_mergers(merge_header_headers) + all_mergers = [] + all_mergers.extend(mergers_yaml) + all_mergers.extend(mergers_header) + if not all_mergers: + all_mergers = DEF_MERGERS + return (payload_yaml, all_mergers) + + def _merge_patch(self, payload): + # JSON doesn't handle comments in this manner, so ensure that + # if we started with this 'type' that we remove it before + # attempting to load it as json (which the jsonpatch library will + # attempt to do). + payload = payload.lstrip() + payload = util.strip_prefix_suffix(payload, prefix=JSONP_PREFIX) + patch = jsonpatch.JsonPatch.from_string(payload) + LOG.debug("Merging by applying json patch %s", patch) + self.cloud_buf = patch.apply(self.cloud_buf, in_place=False) - def _handle_part(self, _data, ctype, filename, payload, _frequency): + def _merge_part(self, payload, headers): + (payload_yaml, my_mergers) = self._extract_mergers(payload, headers) + LOG.debug("Merging by applying %s", my_mergers) + merger = mergers.construct(my_mergers) + self.cloud_buf = merger.merge(self.cloud_buf, payload_yaml) + + def _reset(self): + self.file_names = [] + self.cloud_buf = None + + def handle_part(self, _data, ctype, filename, # pylint: disable=W0221 + payload, _frequency, headers): # pylint: disable=W0613 if ctype == handlers.CONTENT_START: - self.cloud_buf = [] + self._reset() return if ctype == handlers.CONTENT_END: - self._write_cloud_config(self.cloud_buf) - self.cloud_buf = [] + self._write_cloud_config() + self._reset() return - - filename = util.clean_filename(filename) - if not filename: - filename = '??' - self.cloud_buf.extend(["#%s" % (filename), str(payload)]) + try: + # First time through, merge with an empty dict... + if self.cloud_buf is None or not self.file_names: + self.cloud_buf = {} + if ctype == CC_TYPES[JSONP_PREFIX]: + self._merge_patch(payload) + else: + self._merge_part(payload, headers) + # Ensure filename is ok to store + for i in ("\n", "\r", "\t"): + filename = filename.replace(i, " ") + self.file_names.append(filename.strip()) + except: + util.logexc(LOG, "Failed at merging in cloud config part from %s", + filename) diff --git a/cloudinit/handlers/shell_script.py b/cloudinit/handlers/shell_script.py index 6c5c11ca..62289d98 100644 --- a/cloudinit/handlers/shell_script.py +++ b/cloudinit/handlers/shell_script.py @@ -29,6 +29,7 @@ from cloudinit import util from cloudinit.settings import (PER_ALWAYS) LOG = logging.getLogger(__name__) +SHELL_PREFIX = "#!" class ShellScriptPartHandler(handlers.Handler): @@ -38,10 +39,11 @@ class ShellScriptPartHandler(handlers.Handler): def list_types(self): return [ - handlers.type_from_starts_with("#!"), + handlers.type_from_starts_with(SHELL_PREFIX), ] - def _handle_part(self, _data, ctype, filename, payload, _frequency): + def handle_part(self, _data, ctype, filename, # pylint: disable=W0221 + payload, frequency): # pylint: disable=W0613 if ctype in handlers.CONTENT_SIGNALS: # TODO(harlowja): maybe delete existing things here return diff --git a/cloudinit/handlers/upstart_job.py b/cloudinit/handlers/upstart_job.py index 4684f7f2..bac4cad2 100644 --- a/cloudinit/handlers/upstart_job.py +++ b/cloudinit/handlers/upstart_job.py @@ -22,6 +22,7 @@ import os +import re from cloudinit import handlers from cloudinit import log as logging @@ -30,6 +31,7 @@ from cloudinit import util from cloudinit.settings import (PER_INSTANCE) LOG = logging.getLogger(__name__) +UPSTART_PREFIX = "#upstart-job" class UpstartJobPartHandler(handlers.Handler): @@ -39,10 +41,11 @@ class UpstartJobPartHandler(handlers.Handler): def list_types(self): return [ - handlers.type_from_starts_with("#upstart-job"), + handlers.type_from_starts_with(UPSTART_PREFIX), ] - def _handle_part(self, _data, ctype, filename, payload, frequency): + def handle_part(self, _data, ctype, filename, # pylint: disable=W0221 + payload, frequency): if ctype in handlers.CONTENT_SIGNALS: return @@ -65,6 +68,53 @@ class UpstartJobPartHandler(handlers.Handler): path = os.path.join(self.upstart_dir, filename) util.write_file(path, payload, 0644) - # if inotify support is not present in the root filesystem - # (overlayroot) then we need to tell upstart to re-read /etc - util.subp(["initctl", "reload-configuration"], capture=False) + if SUITABLE_UPSTART: + util.subp(["initctl", "reload-configuration"], capture=False) + + +def _has_suitable_upstart(): + # (LP: #1124384) + # a bug in upstart means that invoking reload-configuration + # at this stage in boot causes havoc. So, try to determine if upstart + # is installed, and reloading configuration is OK. + if not os.path.exists("/sbin/initctl"): + return False + try: + (version_out, _err) = util.subp(["initctl", "version"]) + except: + util.logexc(LOG, "initctl version failed") + return False + + # expecting 'initctl version' to output something like: init (upstart X.Y) + if re.match("upstart 1.[0-7][)]", version_out): + return False + if "upstart 0." in version_out: + return False + elif "upstart 1.8" in version_out: + if not os.path.exists("/usr/bin/dpkg-query"): + return False + try: + (dpkg_ver, _err) = util.subp(["dpkg-query", + "--showformat=${Version}", + "--show", "upstart"], rcs=[0, 1]) + except Exception: + util.logexc(LOG, "dpkg-query failed") + return False + + try: + good = "1.8-0ubuntu1.2" + util.subp(["dpkg", "--compare-versions", dpkg_ver, "ge", good]) + return True + except util.ProcessExecutionError as e: + if e.exit_code is 1: + pass + else: + util.logexc(LOG, "dpkg --compare-versions failed [%s]", + e.exit_code) + except Exception as e: + util.logexc(LOG, "dpkg --compare-versions failed") + return False + else: + return True + +SUITABLE_UPSTART = _has_suitable_upstart() diff --git a/cloudinit/helpers.py b/cloudinit/helpers.py index 2077401c..1c46efde 100644 --- a/cloudinit/helpers.py +++ b/cloudinit/helpers.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2012 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # Copyright (C) 2012 Yahoo! Inc. # # Author: Scott Moser <scott.moser@canonical.com> @@ -32,6 +32,7 @@ from cloudinit.settings import (PER_INSTANCE, PER_ALWAYS, PER_ONCE, CFG_ENV_NAME) from cloudinit import log as logging +from cloudinit import type_utils from cloudinit import util LOG = logging.getLogger(__name__) @@ -68,7 +69,7 @@ class FileLock(object): self.fn = fn def __str__(self): - return "<%s using file %r>" % (util.obj_name(self), self.fn) + return "<%s using file %r>" % (type_utils.obj_name(self), self.fn) def canon_sem_name(name): @@ -215,8 +216,8 @@ class ConfigMerger(object): if ds_cfg and isinstance(ds_cfg, (dict)): d_cfgs.append(ds_cfg) except: - util.logexc(LOG, ("Failed loading of datasource" - " config object from %s"), self._ds) + util.logexc(LOG, "Failed loading of datasource config object " + "from %s", self._ds) return d_cfgs def _get_env_configs(self): @@ -226,8 +227,8 @@ class ConfigMerger(object): try: e_cfgs.append(util.read_conf(e_fn)) except: - util.logexc(LOG, ('Failed loading of env. config' - ' from %s'), e_fn) + util.logexc(LOG, 'Failed loading of env. config from %s', + e_fn) return e_cfgs def _get_instance_configs(self): @@ -241,8 +242,8 @@ class ConfigMerger(object): try: i_cfgs.append(util.read_conf(cc_fn)) except: - util.logexc(LOG, ('Failed loading of cloud-config' - ' from %s'), cc_fn) + util.logexc(LOG, 'Failed loading of cloud-config from %s', + cc_fn) return i_cfgs def _read_cfg(self): @@ -258,8 +259,8 @@ class ConfigMerger(object): try: cfgs.append(util.read_conf(c_fn)) except: - util.logexc(LOG, ("Failed loading of configuration" - " from %s"), c_fn) + util.logexc(LOG, "Failed loading of configuration from %s", + c_fn) cfgs.extend(self._get_env_configs()) cfgs.extend(self._get_instance_configs()) @@ -280,6 +281,7 @@ class ContentHandlers(object): def __init__(self): self.registered = {} + self.initialized = [] def __contains__(self, item): return self.is_registered(item) @@ -290,11 +292,13 @@ class ContentHandlers(object): def is_registered(self, content_type): return content_type in self.registered - def register(self, mod): + def register(self, mod, initialized=False): types = set() for t in mod.list_types(): self.registered[t] = mod types.add(t) + if initialized and mod not in self.initialized: + self.initialized.append(mod) return types def _get_handler(self, content_type): diff --git a/cloudinit/log.py b/cloudinit/log.py index da6c2851..622c946c 100644 --- a/cloudinit/log.py +++ b/cloudinit/log.py @@ -44,13 +44,13 @@ NOTSET = logging.NOTSET DEF_CON_FORMAT = '%(asctime)s - %(filename)s[%(levelname)s]: %(message)s' -def setupBasicLogging(): +def setupBasicLogging(level=DEBUG): root = logging.getLogger() console = logging.StreamHandler(sys.stderr) console.setFormatter(logging.Formatter(DEF_CON_FORMAT)) - console.setLevel(DEBUG) + console.setLevel(level) root.addHandler(console) - root.setLevel(DEBUG) + root.setLevel(level) def flushLoggers(root): diff --git a/cloudinit/mergers/__init__.py b/cloudinit/mergers/__init__.py new file mode 100644 index 00000000..0978b2c6 --- /dev/null +++ b/cloudinit/mergers/__init__.py @@ -0,0 +1,167 @@ +# vi: ts=4 expandtab +# +# Copyright (C) 2012 Yahoo! Inc. +# +# Author: Joshua Harlow <harlowja@yahoo-inc.com> +# +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License version 3, as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. + +import re + +from cloudinit import importer +from cloudinit import log as logging +from cloudinit import type_utils + +NAME_MTCH = re.compile(r"(^[a-zA-Z_][A-Za-z0-9_]*)\((.*?)\)$") + +LOG = logging.getLogger(__name__) +DEF_MERGE_TYPE = "list()+dict()+str()" +MERGER_PREFIX = 'm_' +MERGER_ATTR = 'Merger' + + +class UnknownMerger(object): + # Named differently so auto-method finding + # doesn't pick this up if there is ever a type + # named "unknown" + def _handle_unknown(self, _meth_wanted, value, _merge_with): + return value + + # This merging will attempt to look for a '_on_X' method + # in our own object for a given object Y with type X, + # if found it will be called to perform the merge of a source + # object and a object to merge_with. + # + # If not found the merge will be given to a '_handle_unknown' + # function which can decide what to do wit the 2 values. + def merge(self, source, merge_with): + type_name = type_utils.obj_name(source) + type_name = type_name.lower() + method_name = "_on_%s" % (type_name) + meth = None + args = [source, merge_with] + if hasattr(self, method_name): + meth = getattr(self, method_name) + if not meth: + meth = self._handle_unknown + args.insert(0, method_name) + LOG.debug("Merging '%s' into '%s' using method '%s' of '%s'", + type_name, type_utils.obj_name(merge_with), + meth.__name__, self) + return meth(*args) + + +class LookupMerger(UnknownMerger): + def __init__(self, lookups=None): + UnknownMerger.__init__(self) + if lookups is None: + self._lookups = [] + else: + self._lookups = lookups + + def __str__(self): + return 'LookupMerger: (%s)' % (len(self._lookups)) + + # For items which can not be merged by the parent this object + # will lookup in a internally maintained set of objects and + # find which one of those objects can perform the merge. If + # any of the contained objects have the needed method, they + # will be called to perform the merge. + def _handle_unknown(self, meth_wanted, value, merge_with): + meth = None + for merger in self._lookups: + if hasattr(merger, meth_wanted): + # First one that has that method/attr gets to be + # the one that will be called + meth = getattr(merger, meth_wanted) + LOG.debug(("Merging using located merger '%s'" + " since it had method '%s'"), merger, meth_wanted) + break + if not meth: + return UnknownMerger._handle_unknown(self, meth_wanted, + value, merge_with) + return meth(value, merge_with) + + +def dict_extract_mergers(config): + parsed_mergers = [] + raw_mergers = config.pop('merge_how', None) + if raw_mergers is None: + raw_mergers = config.pop('merge_type', None) + if raw_mergers is None: + return parsed_mergers + if isinstance(raw_mergers, (str, basestring)): + return string_extract_mergers(raw_mergers) + for m in raw_mergers: + if isinstance(m, (dict)): + name = m['name'] + name = name.replace("-", "_").strip() + opts = m['settings'] + else: + name = m[0] + if len(m) >= 2: + opts = m[1:] + else: + opts = [] + if name: + parsed_mergers.append((name, opts)) + return parsed_mergers + + +def string_extract_mergers(merge_how): + parsed_mergers = [] + for m_name in merge_how.split("+"): + # Canonicalize the name (so that it can be found + # even when users alter it in various ways) + m_name = m_name.lower().strip() + m_name = m_name.replace("-", "_") + if not m_name: + continue + match = NAME_MTCH.match(m_name) + if not match: + msg = ("Matcher identifer '%s' is not in the right format" % + (m_name)) + raise ValueError(msg) + (m_name, m_ops) = match.groups() + m_ops = m_ops.strip().split(",") + m_ops = [m.strip().lower() for m in m_ops if m.strip()] + parsed_mergers.append((m_name, m_ops)) + return parsed_mergers + + +def default_mergers(): + return tuple(string_extract_mergers(DEF_MERGE_TYPE)) + + +def construct(parsed_mergers): + mergers_to_be = [] + for (m_name, m_ops) in parsed_mergers: + if not m_name.startswith(MERGER_PREFIX): + m_name = MERGER_PREFIX + str(m_name) + merger_locs = importer.find_module(m_name, + [__name__], + [MERGER_ATTR]) + if not merger_locs: + msg = ("Could not find merger module named '%s' " + "with attribute '%s'") % (m_name, MERGER_ATTR) + raise ImportError(msg) + else: + mod = importer.import_module(merger_locs[0]) + mod_attr = getattr(mod, MERGER_ATTR) + mergers_to_be.append((mod_attr, m_ops)) + # Now form them... + mergers = [] + root = LookupMerger(mergers) + for (attr, opts) in mergers_to_be: + mergers.append(attr(root, opts)) + return root diff --git a/cloudinit/mergers/m_dict.py b/cloudinit/mergers/m_dict.py new file mode 100644 index 00000000..a16141fa --- /dev/null +++ b/cloudinit/mergers/m_dict.py @@ -0,0 +1,86 @@ +# vi: ts=4 expandtab +# +# Copyright (C) 2012 Yahoo! Inc. +# +# Author: Joshua Harlow <harlowja@yahoo-inc.com> +# +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License version 3, as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. + +DEF_MERGE_TYPE = 'no_replace' +MERGE_TYPES = ('replace', DEF_MERGE_TYPE,) + + +def _has_any(what, *keys): + for k in keys: + if k in what: + return True + return False + + +class Merger(object): + def __init__(self, merger, opts): + self._merger = merger + # Affects merging behavior... + self._method = DEF_MERGE_TYPE + for m in MERGE_TYPES: + if m in opts: + self._method = m + break + # Affect how recursive merging is done on other primitives. + self._recurse_str = 'recurse_str' in opts + self._recurse_array = _has_any(opts, 'recurse_array', 'recurse_list') + self._allow_delete = 'allow_delete' in opts + # Backwards compat require this to be on. + self._recurse_dict = True + + def __str__(self): + s = ('DictMerger: (method=%s,recurse_str=%s,' + 'recurse_dict=%s,recurse_array=%s,allow_delete=%s)') + s = s % (self._method, self._recurse_str, + self._recurse_dict, self._recurse_array, self._allow_delete) + return s + + def _do_dict_replace(self, value, merge_with, do_replace): + + def merge_same_key(old_v, new_v): + if do_replace: + return new_v + if isinstance(new_v, (list, tuple)) and self._recurse_array: + return self._merger.merge(old_v, new_v) + if isinstance(new_v, (basestring)) and self._recurse_str: + return self._merger.merge(old_v, new_v) + if isinstance(new_v, (dict)) and self._recurse_dict: + return self._merger.merge(old_v, new_v) + # Otherwise leave it be... + return old_v + + for (k, v) in merge_with.items(): + if k in value: + if v is None and self._allow_delete: + value.pop(k) + else: + value[k] = merge_same_key(value[k], v) + else: + value[k] = v + return value + + def _on_dict(self, value, merge_with): + if not isinstance(merge_with, (dict)): + return value + if self._method == 'replace': + merged = self._do_dict_replace(dict(value), merge_with, True) + elif self._method == 'no_replace': + merged = self._do_dict_replace(dict(value), merge_with, False) + else: + raise NotImplementedError("Unknown merge type %s" % (self._method)) + return merged diff --git a/cloudinit/mergers/m_list.py b/cloudinit/mergers/m_list.py new file mode 100644 index 00000000..62999b4e --- /dev/null +++ b/cloudinit/mergers/m_list.py @@ -0,0 +1,87 @@ +# vi: ts=4 expandtab +# +# Copyright (C) 2012 Yahoo! Inc. +# +# Author: Joshua Harlow <harlowja@yahoo-inc.com> +# +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License version 3, as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. + +DEF_MERGE_TYPE = 'replace' +MERGE_TYPES = ('append', 'prepend', DEF_MERGE_TYPE, 'no_replace') + + +def _has_any(what, *keys): + for k in keys: + if k in what: + return True + return False + + +class Merger(object): + def __init__(self, merger, opts): + self._merger = merger + # Affects merging behavior... + self._method = DEF_MERGE_TYPE + for m in MERGE_TYPES: + if m in opts: + self._method = m + break + # Affect how recursive merging is done on other primitives + self._recurse_str = _has_any(opts, 'recurse_str') + self._recurse_dict = _has_any(opts, 'recurse_dict') + self._recurse_array = _has_any(opts, 'recurse_array', 'recurse_list') + + def __str__(self): + return ('ListMerger: (method=%s,recurse_str=%s,' + 'recurse_dict=%s,recurse_array=%s)') % (self._method, + self._recurse_str, + self._recurse_dict, + self._recurse_array) + + def _on_tuple(self, value, merge_with): + return tuple(self._on_list(list(value), merge_with)) + + def _on_list(self, value, merge_with): + if (self._method == 'replace' and + not isinstance(merge_with, (tuple, list))): + return merge_with + + # Ok we now know that what we are merging with is a list or tuple. + merged_list = [] + if self._method == 'prepend': + merged_list.extend(merge_with) + merged_list.extend(value) + return merged_list + elif self._method == 'append': + merged_list.extend(value) + merged_list.extend(merge_with) + return merged_list + + def merge_same_index(old_v, new_v): + if self._method == 'no_replace': + # Leave it be... + return old_v + if isinstance(new_v, (list, tuple)) and self._recurse_array: + return self._merger.merge(old_v, new_v) + if isinstance(new_v, (str, basestring)) and self._recurse_str: + return self._merger.merge(old_v, new_v) + if isinstance(new_v, (dict)) and self._recurse_dict: + return self._merger.merge(old_v, new_v) + return new_v + + # Ok now we are replacing same indexes + merged_list.extend(value) + common_len = min(len(merged_list), len(merge_with)) + for i in xrange(0, common_len): + merged_list[i] = merge_same_index(merged_list[i], merge_with[i]) + return merged_list diff --git a/cloudinit/mergers/m_str.py b/cloudinit/mergers/m_str.py new file mode 100644 index 00000000..e22ce28a --- /dev/null +++ b/cloudinit/mergers/m_str.py @@ -0,0 +1,44 @@ +# -*- coding: utf-8 -*- +# vi: ts=4 expandtab +# +# Copyright (C) 2012 Yahoo! Inc. +# +# Author: Joshua Harlow <harlowja@yahoo-inc.com> +# +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License version 3, as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. + + +class Merger(object): + def __init__(self, _merger, opts): + self._append = 'append' in opts + + def __str__(self): + return 'StringMerger: (append=%s)' % (self._append) + + # On encountering a unicode object to merge value with + # we will for now just proxy into the string method to let it handle it. + def _on_unicode(self, value, merge_with): + return self._on_str(value, merge_with) + + # On encountering a string object to merge with we will + # perform the following action, if appending we will + # merge them together, otherwise we will just return value. + def _on_str(self, value, merge_with): + if not isinstance(value, (basestring)): + return merge_with + if not self._append: + return merge_with + if isinstance(value, unicode): + return value + unicode(merge_with) + else: + return value + str(merge_with) diff --git a/cloudinit/settings.py b/cloudinit/settings.py index 4b95b5b7..5df7f557 100644 --- a/cloudinit/settings.py +++ b/cloudinit/settings.py @@ -32,11 +32,13 @@ CFG_BUILTIN = { 'NoCloud', 'ConfigDrive', 'OpenNebula', + 'Azure', 'AltCloud', 'OVF', 'MAAS', 'Ec2', 'CloudStack', + 'SmartOS', # At the end to act as a 'catch' when none of the above work... 'None', ], diff --git a/cloudinit/sources/DataSourceAltCloud.py b/cloudinit/sources/DataSourceAltCloud.py index 9812bdcb..a834f8eb 100644 --- a/cloudinit/sources/DataSourceAltCloud.py +++ b/cloudinit/sources/DataSourceAltCloud.py @@ -1,10 +1,11 @@ # vi: ts=4 expandtab # # Copyright (C) 2009-2010 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # Copyright (C) 2012 Yahoo! Inc. # # Author: Joe VLcek <JVLcek@RedHat.com> +# Author: Juerg Haefliger <juerg.haefliger@hp.com> # # This program is free software: you can redistribute it and/or modify # it under the terms of the GNU General Public License version 3, as @@ -30,6 +31,7 @@ import os.path from cloudinit import log as logging from cloudinit import sources from cloudinit import util + from cloudinit.util import ProcessExecutionError LOG = logging.getLogger(__name__) @@ -78,7 +80,7 @@ def read_user_data_callback(mount_dir): try: user_data = util.load_file(user_data_file).strip() except IOError: - util.logexc(LOG, ('Failed accessing user data file.')) + util.logexc(LOG, 'Failed accessing user data file.') return None return user_data @@ -91,8 +93,8 @@ class DataSourceAltCloud(sources.DataSource): self.supported_seed_starts = ("/", "file://") def __str__(self): - mstr = "%s [seed=%s]" % (util.obj_name(self), self.seed) - return mstr + root = sources.DataSource.__str__(self) + return "%s [seed=%s]" % (root, self.seed) def get_cloud_type(self): ''' @@ -177,7 +179,7 @@ class DataSourceAltCloud(sources.DataSource): return False # No user data found - util.logexc(LOG, ('Failed accessing user data.')) + util.logexc(LOG, 'Failed accessing user data.') return False def user_data_rhevm(self): @@ -204,12 +206,12 @@ class DataSourceAltCloud(sources.DataSource): (cmd_out, _err) = util.subp(cmd) LOG.debug(('Command: %s\nOutput%s') % (' '.join(cmd), cmd_out)) except ProcessExecutionError, _err: - util.logexc(LOG, (('Failed command: %s\n%s') % \ - (' '.join(cmd), _err.message))) + util.logexc(LOG, 'Failed command: %s\n%s', ' '.join(cmd), + _err.message) return False except OSError, _err: - util.logexc(LOG, (('Failed command: %s\n%s') % \ - (' '.join(cmd), _err.message))) + util.logexc(LOG, 'Failed command: %s\n%s', ' '.join(cmd), + _err.message) return False floppy_dev = '/dev/fd0' @@ -221,12 +223,12 @@ class DataSourceAltCloud(sources.DataSource): (cmd_out, _err) = util.subp(cmd) LOG.debug(('Command: %s\nOutput%s') % (' '.join(cmd), cmd_out)) except ProcessExecutionError, _err: - util.logexc(LOG, (('Failed command: %s\n%s') % \ - (' '.join(cmd), _err.message))) + util.logexc(LOG, 'Failed command: %s\n%s', ' '.join(cmd), + _err.message) return False except OSError, _err: - util.logexc(LOG, (('Failed command: %s\n%s') % \ - (' '.join(cmd), _err.message))) + util.logexc(LOG, 'Failed command: %s\n%s', ' '.join(cmd), + _err.message) return False try: @@ -235,8 +237,8 @@ class DataSourceAltCloud(sources.DataSource): if err.errno != errno.ENOENT: raise except util.MountFailedError: - util.logexc(LOG, ("Failed to mount %s" - " when looking for user data"), floppy_dev) + util.logexc(LOG, "Failed to mount %s when looking for user data", + floppy_dev) self.userdata_raw = return_str self.metadata = META_DATA_NOT_SUPPORTED @@ -271,8 +273,8 @@ class DataSourceAltCloud(sources.DataSource): if err.errno != errno.ENOENT: raise except util.MountFailedError: - util.logexc(LOG, ("Failed to mount %s" - " when looking for user data"), cdrom_dev) + util.logexc(LOG, "Failed to mount %s when looking for user " + "data", cdrom_dev) self.userdata_raw = return_str self.metadata = META_DATA_NOT_SUPPORTED diff --git a/cloudinit/sources/DataSourceAzure.py b/cloudinit/sources/DataSourceAzure.py new file mode 100644 index 00000000..66d7728b --- /dev/null +++ b/cloudinit/sources/DataSourceAzure.py @@ -0,0 +1,502 @@ +# vi: ts=4 expandtab +# +# Copyright (C) 2013 Canonical Ltd. +# +# Author: Scott Moser <scott.moser@canonical.com> +# +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License version 3, as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. + +import base64 +import crypt +import os +import os.path +import time +from xml.dom import minidom + +from cloudinit import log as logging +from cloudinit import sources +from cloudinit import util + +LOG = logging.getLogger(__name__) + +DS_NAME = 'Azure' +DEFAULT_METADATA = {"instance-id": "iid-AZURE-NODE"} +AGENT_START = ['service', 'walinuxagent', 'start'] +BOUNCE_COMMAND = ['sh', '-xc', + "i=$interface; x=0; ifdown $i || x=$?; ifup $i || x=$?; exit $x"] + +BUILTIN_DS_CONFIG = { + 'agent_command': AGENT_START, + 'data_dir': "/var/lib/waagent", + 'set_hostname': True, + 'hostname_bounce': { + 'interface': 'eth0', + 'policy': True, + 'command': BOUNCE_COMMAND, + 'hostname_command': 'hostname', + } +} +DS_CFG_PATH = ['datasource', DS_NAME] + + +class DataSourceAzureNet(sources.DataSource): + def __init__(self, sys_cfg, distro, paths): + sources.DataSource.__init__(self, sys_cfg, distro, paths) + self.seed_dir = os.path.join(paths.seed_dir, 'azure') + self.cfg = {} + self.seed = None + self.ds_cfg = util.mergemanydict([ + util.get_cfg_by_path(sys_cfg, DS_CFG_PATH, {}), + BUILTIN_DS_CONFIG]) + + def __str__(self): + root = sources.DataSource.__str__(self) + return "%s [seed=%s]" % (root, self.seed) + + def get_data(self): + # azure removes/ejects the cdrom containing the ovf-env.xml + # file on reboot. So, in order to successfully reboot we + # need to look in the datadir and consider that valid + ddir = self.ds_cfg['data_dir'] + + candidates = [self.seed_dir] + candidates.extend(list_possible_azure_ds_devs()) + if ddir: + candidates.append(ddir) + + found = None + + for cdev in candidates: + try: + if cdev.startswith("/dev/"): + ret = util.mount_cb(cdev, load_azure_ds_dir) + else: + ret = load_azure_ds_dir(cdev) + + except NonAzureDataSource: + continue + except BrokenAzureDataSource as exc: + raise exc + except util.MountFailedError: + LOG.warn("%s was not mountable" % cdev) + continue + + (md, self.userdata_raw, cfg, files) = ret + self.seed = cdev + self.metadata = util.mergemanydict([md, DEFAULT_METADATA]) + self.cfg = cfg + found = cdev + + LOG.debug("found datasource in %s", cdev) + break + + if not found: + return False + + if found == ddir: + LOG.debug("using files cached in %s", ddir) + + # now update ds_cfg to reflect contents pass in config + usercfg = util.get_cfg_by_path(self.cfg, DS_CFG_PATH, {}) + self.ds_cfg = util.mergemanydict([usercfg, self.ds_cfg]) + mycfg = self.ds_cfg + + # walinux agent writes files world readable, but expects + # the directory to be protected. + write_files(mycfg['data_dir'], files, dirmode=0700) + + # handle the hostname 'publishing' + try: + handle_set_hostname(mycfg.get('set_hostname'), + self.metadata.get('local-hostname'), + mycfg['hostname_bounce']) + except Exception as e: + LOG.warn("Failed publishing hostname: %s" % e) + util.logexc(LOG, "handling set_hostname failed") + + try: + invoke_agent(mycfg['agent_command']) + except util.ProcessExecutionError: + # claim the datasource even if the command failed + util.logexc(LOG, "agent command '%s' failed.", + mycfg['agent_command']) + + shcfgxml = os.path.join(mycfg['data_dir'], "SharedConfig.xml") + wait_for = [shcfgxml] + + fp_files = [] + for pk in self.cfg.get('_pubkeys', []): + bname = pk['fingerprint'] + ".crt" + fp_files += [os.path.join(mycfg['data_dir'], bname)] + + missing = util.log_time(logfunc=LOG.debug, msg="waiting for files", + func=wait_for_files, + args=(wait_for + fp_files,)) + if len(missing): + LOG.warn("Did not find files, but going on: %s", missing) + + if shcfgxml in missing: + LOG.warn("SharedConfig.xml missing, using static instance-id") + else: + try: + self.metadata['instance-id'] = iid_from_shared_config(shcfgxml) + except ValueError as e: + LOG.warn("failed to get instance id in %s: %s" % (shcfgxml, e)) + + pubkeys = pubkeys_from_crt_files(fp_files) + + self.metadata['public-keys'] = pubkeys + + return True + + def get_config_obj(self): + return self.cfg + + +def handle_set_hostname(enabled, hostname, cfg): + if not util.is_true(enabled): + return + + if not hostname: + LOG.warn("set_hostname was true but no local-hostname") + return + + apply_hostname_bounce(hostname=hostname, policy=cfg['policy'], + interface=cfg['interface'], + command=cfg['command'], + hostname_command=cfg['hostname_command']) + + +def apply_hostname_bounce(hostname, policy, interface, command, + hostname_command="hostname"): + # set the hostname to 'hostname' if it is not already set to that. + # then, if policy is not off, bounce the interface using command + prev_hostname = util.subp(hostname_command, capture=True)[0].strip() + + util.subp([hostname_command, hostname]) + + msg = ("phostname=%s hostname=%s policy=%s interface=%s" % + (prev_hostname, hostname, policy, interface)) + + if util.is_false(policy): + LOG.debug("pubhname: policy false, skipping [%s]", msg) + return + + if prev_hostname == hostname and policy != "force": + LOG.debug("pubhname: no change, policy != force. skipping. [%s]", msg) + return + + env = os.environ.copy() + env['interface'] = interface + env['hostname'] = hostname + env['old_hostname'] = prev_hostname + + if command == "builtin": + command = BOUNCE_COMMAND + + LOG.debug("pubhname: publishing hostname [%s]", msg) + shell = not isinstance(command, (list, tuple)) + # capture=False, see comments in bug 1202758 and bug 1206164. + util.log_time(logfunc=LOG.debug, msg="publishing hostname", + get_uptime=True, func=util.subp, + kwargs={'args': command, 'shell': shell, 'capture': False, + 'env': env}) + + +def crtfile_to_pubkey(fname): + pipeline = ('openssl x509 -noout -pubkey < "$0" |' + 'ssh-keygen -i -m PKCS8 -f /dev/stdin') + (out, _err) = util.subp(['sh', '-c', pipeline, fname], capture=True) + return out.rstrip() + + +def pubkeys_from_crt_files(flist): + pubkeys = [] + errors = [] + for fname in flist: + try: + pubkeys.append(crtfile_to_pubkey(fname)) + except util.ProcessExecutionError: + errors.extend(fname) + + if errors: + LOG.warn("failed to convert the crt files to pubkey: %s" % errors) + + return pubkeys + + +def wait_for_files(flist, maxwait=60, naplen=.5): + need = set(flist) + waited = 0 + while waited < maxwait: + need -= set([f for f in need if os.path.exists(f)]) + if len(need) == 0: + return [] + time.sleep(naplen) + waited += naplen + return need + + +def write_files(datadir, files, dirmode=None): + if not datadir: + return + if not files: + files = {} + util.ensure_dir(datadir, dirmode) + for (name, content) in files.items(): + util.write_file(filename=os.path.join(datadir, name), + content=content, mode=0600) + + +def invoke_agent(cmd): + # this is a function itself to simplify patching it for test + if cmd: + LOG.debug("invoking agent: %s" % cmd) + util.subp(cmd, shell=(not isinstance(cmd, list))) + else: + LOG.debug("not invoking agent") + + +def find_child(node, filter_func): + ret = [] + if not node.hasChildNodes(): + return ret + for child in node.childNodes: + if filter_func(child): + ret.append(child) + return ret + + +def load_azure_ovf_pubkeys(sshnode): + # This parses a 'SSH' node formatted like below, and returns + # an array of dicts. + # [{'fp': '6BE7A7C3C8A8F4B123CCA5D0C2F1BE4CA7B63ED7', + # 'path': 'where/to/go'}] + # + # <SSH><PublicKeys> + # <PublicKey><Fingerprint>ABC</FingerPrint><Path>/ABC</Path> + # ... + # </PublicKeys></SSH> + results = find_child(sshnode, lambda n: n.localName == "PublicKeys") + if len(results) == 0: + return [] + if len(results) > 1: + raise BrokenAzureDataSource("Multiple 'PublicKeys'(%s) in SSH node" % + len(results)) + + pubkeys_node = results[0] + pubkeys = find_child(pubkeys_node, lambda n: n.localName == "PublicKey") + + if len(pubkeys) == 0: + return [] + + found = [] + text_node = minidom.Document.TEXT_NODE + + for pk_node in pubkeys: + if not pk_node.hasChildNodes(): + continue + cur = {'fingerprint': "", 'path': ""} + for child in pk_node.childNodes: + if (child.nodeType == text_node or not child.localName): + continue + + name = child.localName.lower() + + if name not in cur.keys(): + continue + + if (len(child.childNodes) != 1 or + child.childNodes[0].nodeType != text_node): + continue + + cur[name] = child.childNodes[0].wholeText.strip() + found.append(cur) + + return found + + +def single_node_at_path(node, pathlist): + curnode = node + for tok in pathlist: + results = find_child(curnode, lambda n: n.localName == tok) + if len(results) == 0: + raise ValueError("missing %s token in %s" % (tok, str(pathlist))) + if len(results) > 1: + raise ValueError("found %s nodes of type %s looking for %s" % + (len(results), tok, str(pathlist))) + curnode = results[0] + + return curnode + + +def read_azure_ovf(contents): + try: + dom = minidom.parseString(contents) + except Exception as e: + raise NonAzureDataSource("invalid xml: %s" % e) + + results = find_child(dom.documentElement, + lambda n: n.localName == "ProvisioningSection") + + if len(results) == 0: + raise NonAzureDataSource("No ProvisioningSection") + if len(results) > 1: + raise BrokenAzureDataSource("found '%d' ProvisioningSection items" % + len(results)) + provSection = results[0] + + lpcs_nodes = find_child(provSection, + lambda n: n.localName == "LinuxProvisioningConfigurationSet") + + if len(results) == 0: + raise NonAzureDataSource("No LinuxProvisioningConfigurationSet") + if len(results) > 1: + raise BrokenAzureDataSource("found '%d' %ss" % + ("LinuxProvisioningConfigurationSet", + len(results))) + lpcs = lpcs_nodes[0] + + if not lpcs.hasChildNodes(): + raise BrokenAzureDataSource("no child nodes of configuration set") + + md_props = 'seedfrom' + md = {'azure_data': {}} + cfg = {} + ud = "" + password = None + username = None + + for child in lpcs.childNodes: + if child.nodeType == dom.TEXT_NODE or not child.localName: + continue + + name = child.localName.lower() + + simple = False + value = "" + if (len(child.childNodes) == 1 and + child.childNodes[0].nodeType == dom.TEXT_NODE): + simple = True + value = child.childNodes[0].wholeText + + attrs = {k: v for k, v in child.attributes.items()} + + # we accept either UserData or CustomData. If both are present + # then behavior is undefined. + if (name == "userdata" or name == "customdata"): + if attrs.get('encoding') in (None, "base64"): + ud = base64.b64decode(''.join(value.split())) + else: + ud = value + elif name == "username": + username = value + elif name == "userpassword": + password = value + elif name == "hostname": + md['local-hostname'] = value + elif name == "dscfg": + if attrs.get('encoding') in (None, "base64"): + dscfg = base64.b64decode(''.join(value.split())) + else: + dscfg = value + cfg['datasource'] = {DS_NAME: util.load_yaml(dscfg, default={})} + elif name == "ssh": + cfg['_pubkeys'] = load_azure_ovf_pubkeys(child) + elif name == "disablesshpasswordauthentication": + cfg['ssh_pwauth'] = util.is_false(value) + elif simple: + if name in md_props: + md[name] = value + else: + md['azure_data'][name] = value + + defuser = {} + if username: + defuser['name'] = username + if password: + defuser['passwd'] = encrypt_pass(password) + defuser['lock_passwd'] = False + + if defuser: + cfg['system_info'] = {'default_user': defuser} + + if 'ssh_pwauth' not in cfg and password: + cfg['ssh_pwauth'] = True + + return (md, ud, cfg) + + +def encrypt_pass(password, salt_id="$6$"): + return crypt.crypt(password, salt_id + util.rand_str(strlen=16)) + + +def list_possible_azure_ds_devs(): + # return a sorted list of devices that might have a azure datasource + devlist = [] + for fstype in ("iso9660", "udf"): + devlist.extend(util.find_devs_with("TYPE=%s" % fstype)) + + devlist.sort(reverse=True) + return devlist + + +def load_azure_ds_dir(source_dir): + ovf_file = os.path.join(source_dir, "ovf-env.xml") + + if not os.path.isfile(ovf_file): + raise NonAzureDataSource("No ovf-env file found") + + with open(ovf_file, "r") as fp: + contents = fp.read() + + md, ud, cfg = read_azure_ovf(contents) + return (md, ud, cfg, {'ovf-env.xml': contents}) + + +def iid_from_shared_config(path): + with open(path, "rb") as fp: + content = fp.read() + return iid_from_shared_config_content(content) + + +def iid_from_shared_config_content(content): + """ + find INSTANCE_ID in: + <?xml version="1.0" encoding="utf-8"?> + <SharedConfig version="1.0.0.0" goalStateIncarnation="1"> + <Deployment name="INSTANCE_ID" guid="{...}" incarnation="0"> + <Service name="..." guid="{00000000-0000-0000-0000-000000000000}" /> + """ + dom = minidom.parseString(content) + depnode = single_node_at_path(dom, ["SharedConfig", "Deployment"]) + return depnode.attributes.get('name').value + + +class BrokenAzureDataSource(Exception): + pass + + +class NonAzureDataSource(Exception): + pass + + +# Used to match classes to dependencies +datasources = [ + (DataSourceAzureNet, (sources.DEP_FILESYSTEM, sources.DEP_NETWORK)), +] + + +# Return a list of data sources that match this set of dependencies +def get_datasource_list(depends): + return sources.list_from_depends(depends, datasources) diff --git a/cloudinit/sources/DataSourceCloudStack.py b/cloudinit/sources/DataSourceCloudStack.py index 275caf0d..08f661e4 100644 --- a/cloudinit/sources/DataSourceCloudStack.py +++ b/cloudinit/sources/DataSourceCloudStack.py @@ -4,11 +4,13 @@ # Copyright (C) 2012 Cosmin Luta # Copyright (C) 2012 Yahoo! Inc. # Copyright (C) 2012 Gerard Dethier +# Copyright (C) 2013 Hewlett-Packard Development Company, L.P. # # Author: Cosmin Luta <q4break@gmail.com> # Author: Scott Moser <scott.moser@canonical.com> # Author: Joshua Harlow <harlowja@yahoo-inc.com> # Author: Gerard Dethier <g.dethier@gmail.com> +# Author: Juerg Haefliger <juerg.haefliger@hp.com> # # This program is free software: you can redistribute it and/or modify # it under the terms of the GNU General Public License version 3, as @@ -48,9 +50,6 @@ class DataSourceCloudStack(sources.DataSource): raise RuntimeError("No virtual router found!") self.metadata_address = "http://%s/" % (vr_addr) - def __str__(self): - return util.obj_name(self) - def _get_url_settings(self): mcfg = self.ds_cfg if not mcfg: @@ -112,8 +111,8 @@ class DataSourceCloudStack(sources.DataSource): int(time.time() - start_time)) return True except Exception: - util.logexc(LOG, ('Failed fetching from metadata ' - 'service %s'), self.metadata_address) + util.logexc(LOG, 'Failed fetching from metadata service %s', + self.metadata_address) return False def get_instance_id(self): diff --git a/cloudinit/sources/DataSourceConfigDrive.py b/cloudinit/sources/DataSourceConfigDrive.py index ec016a1d..835f2a9a 100644 --- a/cloudinit/sources/DataSourceConfigDrive.py +++ b/cloudinit/sources/DataSourceConfigDrive.py @@ -51,7 +51,9 @@ class DataSourceConfigDrive(sources.DataSource): self.ec2_metadata = None def __str__(self): - mstr = "%s [%s,ver=%s]" % (util.obj_name(self), self.dsmode, + root = sources.DataSource.__str__(self) + mstr = "%s [%s,ver=%s]" % (root, + self.dsmode, self.version) mstr += "[source=%s]" % (self.source) return mstr @@ -152,7 +154,7 @@ class DataSourceConfigDrive(sources.DataSource): return False md = results['metadata'] - md = util.mergedict(md, DEFAULT_METADATA) + md = util.mergemanydict([md, DEFAULT_METADATA]) # Perform some metadata 'fixups' # @@ -256,6 +258,10 @@ def find_candidate_devs(): * labeled with 'config-2' """ + # Query optical drive to get it in blkid cache for 2.6 kernels + util.find_devs_with(path="/dev/sr0") + util.find_devs_with(path="/dev/sr1") + by_fstype = (util.find_devs_with("TYPE=vfat") + util.find_devs_with("TYPE=iso9660")) by_label = util.find_devs_with("LABEL=config-2") diff --git a/cloudinit/sources/DataSourceEc2.py b/cloudinit/sources/DataSourceEc2.py index 2db53446..f010e640 100644 --- a/cloudinit/sources/DataSourceEc2.py +++ b/cloudinit/sources/DataSourceEc2.py @@ -49,9 +49,6 @@ class DataSourceEc2(sources.DataSource): self.seed_dir = os.path.join(paths.seed_dir, "ec2") self.api_ver = DEF_MD_VERSION - def __str__(self): - return util.obj_name(self) - def get_data(self): seed_ret = {} if util.read_optional_seed(seed_ret, base=(self.seed_dir + "/")): diff --git a/cloudinit/sources/DataSourceMAAS.py b/cloudinit/sources/DataSourceMAAS.py index b55d8a21..dfe90bc6 100644 --- a/cloudinit/sources/DataSourceMAAS.py +++ b/cloudinit/sources/DataSourceMAAS.py @@ -27,7 +27,7 @@ import urllib2 from cloudinit import log as logging from cloudinit import sources -from cloudinit import url_helper as uhelp +from cloudinit import url_helper from cloudinit import util LOG = logging.getLogger(__name__) @@ -50,7 +50,8 @@ class DataSourceMAAS(sources.DataSource): self.oauth_clockskew = None def __str__(self): - return "%s [%s]" % (util.obj_name(self), self.base_url) + root = sources.DataSource.__str__(self) + return "%s [%s]" % (root, self.base_url) def get_data(self): mcfg = self.ds_cfg @@ -79,7 +80,8 @@ class DataSourceMAAS(sources.DataSource): self.base_url = url (userdata, metadata) = read_maas_seed_url(self.base_url, - self.md_headers) + self._md_headers, + paths=self.paths) self.userdata_raw = userdata self.metadata = metadata return True @@ -87,7 +89,7 @@ class DataSourceMAAS(sources.DataSource): util.logexc(LOG, "Failed fetching metadata from url %s", url) return False - def md_headers(self, url): + def _md_headers(self, url): mcfg = self.ds_cfg # If we are missing token_key, token_secret or consumer_key @@ -131,36 +133,37 @@ class DataSourceMAAS(sources.DataSource): starttime = time.time() check_url = "%s/%s/meta-data/instance-id" % (url, MD_VERSION) urls = [check_url] - url = uhelp.wait_for_url(urls=urls, max_wait=max_wait, - timeout=timeout, exception_cb=self._except_cb, - headers_cb=self.md_headers) + url = url_helper.wait_for_url(urls=urls, max_wait=max_wait, + timeout=timeout, + exception_cb=self._except_cb, + headers_cb=self._md_headers) if url: LOG.debug("Using metadata source: '%s'", url) else: LOG.critical("Giving up on md from %s after %i seconds", - urls, int(time.time() - starttime)) + urls, int(time.time() - starttime)) return bool(url) def _except_cb(self, msg, exception): - if not (isinstance(exception, urllib2.HTTPError) and + if not (isinstance(exception, url_helper.UrlError) and (exception.code == 403 or exception.code == 401)): return + if 'date' not in exception.headers: - LOG.warn("date field not in %d headers" % exception.code) + LOG.warn("Missing header 'date' in %s response", exception.code) return date = exception.headers['date'] - try: ret_time = time.mktime(parsedate(date)) - except: - LOG.warn("failed to convert datetime '%s'") + except Exception as e: + LOG.warn("Failed to convert datetime '%s': %s", date, e) return self.oauth_clockskew = int(ret_time - time.time()) - LOG.warn("set oauth clockskew to %d" % self.oauth_clockskew) + LOG.warn("Setting oauth clockskew to %d", self.oauth_clockskew) return @@ -188,11 +191,11 @@ def read_maas_seed_dir(seed_d): def read_maas_seed_url(seed_url, header_cb=None, timeout=None, - version=MD_VERSION): + version=MD_VERSION, paths=None): """ Read the maas datasource at seed_url. - header_cb is a method that should return a headers dictionary that will - be given to urllib2.Request() + - header_cb is a method that should return a headers dictionary for + a given url Expected format of seed_url is are the following files: * <seed_url>/<version>/meta-data/instance-id @@ -215,18 +218,28 @@ def read_maas_seed_url(seed_url, header_cb=None, timeout=None, md = {} for name in file_order: url = files.get(name) - if header_cb: - headers = header_cb(url) + if not header_cb: + def _cb(url): + return {} + header_cb = _cb + + if name == 'user-data': + retries = 0 else: - headers = {} + retries = None + try: - resp = uhelp.readurl(url, headers=headers, timeout=timeout) + ssl_details = util.fetch_ssl_details(paths) + resp = util.read_file_or_url(url, retries=retries, + headers_cb=header_cb, + timeout=timeout, + ssl_details=ssl_details) if resp.ok(): md[name] = str(resp) else: LOG.warn(("Fetching from %s resulted in" " an invalid http code %s"), url, resp.code) - except urllib2.HTTPError as e: + except url_helper.UrlError as e: if e.code != 404: raise return check_seed_contents(md, seed_url) @@ -369,7 +382,8 @@ if __name__ == "__main__": if args.subcmd == "check-seed": if args.url.startswith("http"): (userdata, metadata) = read_maas_seed_url(args.url, - header_cb=my_headers, version=args.apiver) + header_cb=my_headers, + version=args.apiver) else: (userdata, metadata) = read_maas_seed_url(args.url) print "=== userdata ===" diff --git a/cloudinit/sources/DataSourceNoCloud.py b/cloudinit/sources/DataSourceNoCloud.py index 097bbc52..4ef92a56 100644 --- a/cloudinit/sources/DataSourceNoCloud.py +++ b/cloudinit/sources/DataSourceNoCloud.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2009-2010 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # Copyright (C) 2012 Yahoo! Inc. # # Author: Scott Moser <scott.moser@canonical.com> @@ -40,9 +40,8 @@ class DataSourceNoCloud(sources.DataSource): self.supported_seed_starts = ("/", "file://") def __str__(self): - mstr = "%s [seed=%s][dsmode=%s]" % (util.obj_name(self), - self.seed, self.dsmode) - return mstr + root = sources.DataSource.__str__(self) + return "%s [seed=%s][dsmode=%s]" % (root, self.seed, self.dsmode) def get_data(self): defaults = { @@ -65,7 +64,7 @@ class DataSourceNoCloud(sources.DataSource): # Check to see if the seed dir has data. seedret = {} if util.read_optional_seed(seedret, base=self.seed_dir + "/"): - md = util.mergedict(md, seedret['meta-data']) + md = util.mergemanydict([md, seedret['meta-data']]) ud = seedret['user-data'] found.append(self.seed_dir) LOG.debug("Using seeded cache data from %s", self.seed_dir) @@ -82,15 +81,19 @@ class DataSourceNoCloud(sources.DataSource): if self.ds_cfg['user-data']: ud = self.ds_cfg['user-data'] if self.ds_cfg['meta-data'] is not False: - md = util.mergedict(md, self.ds_cfg['meta-data']) + md = util.mergemanydict([md, self.ds_cfg['meta-data']]) if 'ds_config' not in found: found.append("ds_config") - if self.ds_cfg.get('fs_label', "cidata"): + label = self.ds_cfg.get('fs_label', "cidata") + if label is not None: + # Query optical drive to get it in blkid cache for 2.6 kernels + util.find_devs_with(path="/dev/sr0") + util.find_devs_with(path="/dev/sr1") + fslist = util.find_devs_with("TYPE=vfat") fslist.extend(util.find_devs_with("TYPE=iso9660")) - label = self.ds_cfg.get('fs_label') label_list = util.find_devs_with("LABEL=%s" % label) devlist = list(set(fslist) & set(label_list)) devlist.sort(reverse=True) @@ -100,7 +103,7 @@ class DataSourceNoCloud(sources.DataSource): LOG.debug("Attempting to use data from %s", dev) (newmd, newud) = util.mount_cb(dev, util.read_seeded) - md = util.mergedict(newmd, md) + md = util.mergemanydict([newmd, md]) ud = newud # For seed from a device, the default mode is 'net'. @@ -116,8 +119,8 @@ class DataSourceNoCloud(sources.DataSource): if e.errno != errno.ENOENT: raise except util.MountFailedError: - util.logexc(LOG, ("Failed to mount %s" - " when looking for data"), dev) + util.logexc(LOG, "Failed to mount %s when looking for " + "data", dev) # There was no indication on kernel cmdline or data # in the seeddir suggesting this handler should be used. @@ -150,11 +153,11 @@ class DataSourceNoCloud(sources.DataSource): LOG.debug("Using seeded cache data from %s", seedfrom) # Values in the command line override those from the seed - md = util.mergedict(md, md_seed) + md = util.mergemanydict([md, md_seed]) found.append(seedfrom) # Now that we have exhausted any other places merge in the defaults - md = util.mergedict(md, defaults) + md = util.mergemanydict([md, defaults]) # Update the network-interfaces if metadata had 'network-interfaces' # entry and this is the local datasource, or 'seedfrom' was used diff --git a/cloudinit/sources/DataSourceNone.py b/cloudinit/sources/DataSourceNone.py index c2125bee..12a8a992 100644 --- a/cloudinit/sources/DataSourceNone.py +++ b/cloudinit/sources/DataSourceNone.py @@ -18,7 +18,6 @@ from cloudinit import log as logging from cloudinit import sources -from cloudinit import util LOG = logging.getLogger(__name__) @@ -41,9 +40,6 @@ class DataSourceNone(sources.DataSource): def get_instance_id(self): return 'iid-datasource-none' - def __str__(self): - return util.obj_name(self) - @property def is_disconnected(self): return True diff --git a/cloudinit/sources/DataSourceOVF.py b/cloudinit/sources/DataSourceOVF.py index e90150c6..77b43e17 100644 --- a/cloudinit/sources/DataSourceOVF.py +++ b/cloudinit/sources/DataSourceOVF.py @@ -43,7 +43,8 @@ class DataSourceOVF(sources.DataSource): self.supported_seed_starts = ("/", "file://") def __str__(self): - return "%s [seed=%s]" % (util.obj_name(self), self.seed) + root = sources.DataSource.__str__(self) + return "%s [seed=%s]" % (root, self.seed) def get_data(self): found = [] @@ -93,11 +94,11 @@ class DataSourceOVF(sources.DataSource): (md_seed, ud) = util.read_seeded(seedfrom, timeout=None) LOG.debug("Using seeded cache data from %s", seedfrom) - md = util.mergedict(md, md_seed) + md = util.mergemanydict([md, md_seed]) found.append(seedfrom) # Now that we have exhausted any other places merge in the defaults - md = util.mergedict(md, defaults) + md = util.mergemanydict([md, defaults]) self.seed = ",".join(found) self.metadata = md @@ -193,6 +194,11 @@ def transport_iso9660(require_iso=True): if contents is not False: return (contents, dev, fname) + if require_iso: + mtype = "iso9660" + else: + mtype = None + devs = os.listdir("/dev/") devs.sort() for dev in devs: @@ -210,7 +216,7 @@ def transport_iso9660(require_iso=True): try: (fname, contents) = util.mount_cb(fullp, - get_ovf_env, mtype="iso9660") + get_ovf_env, mtype=mtype) except util.MountFailedError: LOG.debug("%s not mountable as iso9660" % fullp) continue diff --git a/cloudinit/sources/DataSourceSmartOS.py b/cloudinit/sources/DataSourceSmartOS.py new file mode 100644 index 00000000..d348d20b --- /dev/null +++ b/cloudinit/sources/DataSourceSmartOS.py @@ -0,0 +1,244 @@ +# vi: ts=4 expandtab +# +# Copyright (C) 2013 Canonical Ltd. +# +# Author: Ben Howard <ben.howard@canonical.com> +# +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License version 3, as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. +# +# +# Datasource for provisioning on SmartOS. This works on Joyent +# and public/private Clouds using SmartOS. +# +# SmartOS hosts use a serial console (/dev/ttyS1) on Linux Guests. +# The meta-data is transmitted via key/value pairs made by +# requests on the console. For example, to get the hostname, you +# would send "GET hostname" on /dev/ttyS1. +# + + +import base64 +from cloudinit import log as logging +from cloudinit import sources +from cloudinit import util +import os +import os.path +import serial + +DEF_TTY_LOC = '/dev/ttyS1' +DEF_TTY_TIMEOUT = 60 +LOG = logging.getLogger(__name__) + +SMARTOS_ATTRIB_MAP = { + #Cloud-init Key : (SmartOS Key, Strip line endings) + 'local-hostname': ('hostname', True), + 'public-keys': ('root_authorized_keys', True), + 'user-script': ('user-script', False), + 'user-data': ('user-data', False), + 'iptables_disable': ('iptables_disable', True), + 'motd_sys_info': ('motd_sys_info', True), +} + +# These are values which will never be base64 encoded. +# They come from the cloud platform, not user +SMARTOS_NO_BASE64 = ['root_authorized_keys', 'motd_sys_info', + 'iptables_disable'] + + +class DataSourceSmartOS(sources.DataSource): + def __init__(self, sys_cfg, distro, paths): + sources.DataSource.__init__(self, sys_cfg, distro, paths) + self.seed_dir = os.path.join(paths.seed_dir, 'sdc') + self.is_smartdc = None + + self.seed = self.ds_cfg.get("serial_device", DEF_TTY_LOC) + self.seed_timeout = self.ds_cfg.get("serial_timeout", DEF_TTY_TIMEOUT) + self.smartos_no_base64 = self.ds_cfg.get('no_base64_decode', + SMARTOS_NO_BASE64) + self.b64_keys = self.ds_cfg.get('base64_keys', []) + self.b64_all = self.ds_cfg.get('base64_all', False) + + def __str__(self): + root = sources.DataSource.__str__(self) + return "%s [seed=%s]" % (root, self.seed) + + def get_data(self): + md = {} + ud = "" + + if not os.path.exists(self.seed): + LOG.debug("Host does not appear to be on SmartOS") + return False + self.seed = self.seed + + dmi_info = dmi_data() + if dmi_info is False: + LOG.debug("No dmidata utility found") + return False + + system_uuid, system_type = dmi_info + if 'smartdc' not in system_type.lower(): + LOG.debug("Host is not on SmartOS. system_type=%s", system_type) + return False + self.is_smartdc = True + md['instance-id'] = system_uuid + + b64_keys = self.query('base64_keys', strip=True, b64=False) + if b64_keys is not None: + self.b64_keys = [k.strip() for k in str(b64_keys).split(',')] + + b64_all = self.query('base64_all', strip=True, b64=False) + if b64_all is not None: + self.b64_all = util.is_true(b64_all) + + for ci_noun, attribute in SMARTOS_ATTRIB_MAP.iteritems(): + smartos_noun, strip = attribute + md[ci_noun] = self.query(smartos_noun, strip=strip) + + if not md['local-hostname']: + md['local-hostname'] = system_uuid + + ud = None + if md['user-data']: + ud = md['user-data'] + elif md['user-script']: + ud = md['user-script'] + + self.metadata = md + self.userdata_raw = ud + return True + + def get_instance_id(self): + return self.metadata['instance-id'] + + def query(self, noun, strip=False, default=None, b64=None): + if b64 is None: + if noun in self.smartos_no_base64: + b64 = False + elif self.b64_all or noun in self.b64_keys: + b64 = True + + return query_data(noun=noun, strip=strip, seed_device=self.seed, + seed_timeout=self.seed_timeout, default=default, + b64=b64) + + +def get_serial(seed_device, seed_timeout): + """This is replaced in unit testing, allowing us to replace + serial.Serial with a mocked class. + + The timeout value of 60 seconds should never be hit. The value + is taken from SmartOS own provisioning tools. Since we are reading + each line individually up until the single ".", the transfer is + usually very fast (i.e. microseconds) to get the response. + """ + if not seed_device: + raise AttributeError("seed_device value is not set") + + ser = serial.Serial(seed_device, timeout=seed_timeout) + if not ser.isOpen(): + raise SystemError("Unable to open %s" % seed_device) + + return ser + + +def query_data(noun, seed_device, seed_timeout, strip=False, default=None, + b64=None): + """Makes a request to via the serial console via "GET <NOUN>" + + In the response, the first line is the status, while subsequent lines + are is the value. A blank line with a "." is used to indicate end of + response. + + If the response is expected to be base64 encoded, then set b64encoded + to true. Unfortantely, there is no way to know if something is 100% + encoded, so this method relies on being told if the data is base64 or + not. + """ + + if not noun: + return False + + ser = get_serial(seed_device, seed_timeout) + ser.write("GET %s\n" % noun.rstrip()) + status = str(ser.readline()).rstrip() + response = [] + eom_found = False + + if 'SUCCESS' not in status: + ser.close() + return default + + while not eom_found: + m = ser.readline() + if m.rstrip() == ".": + eom_found = True + else: + response.append(m) + + ser.close() + + if b64 is None: + b64 = query_data('b64-%s' % noun, seed_device=seed_device, + seed_timeout=seed_timeout, b64=False, + default=False, strip=True) + b64 = util.is_true(b64) + + resp = None + if b64 or strip: + resp = "".join(response).rstrip() + else: + resp = "".join(response) + + if b64: + try: + return base64.b64decode(resp) + except TypeError: + LOG.warn("Failed base64 decoding key '%s'", noun) + return resp + + return resp + + +def dmi_data(): + sys_uuid, sys_type = None, None + dmidecode_path = util.which('dmidecode') + if not dmidecode_path: + return False + + sys_uuid_cmd = [dmidecode_path, "-s", "system-uuid"] + try: + LOG.debug("Getting hostname from dmidecode") + (sys_uuid, _err) = util.subp(sys_uuid_cmd) + except Exception as e: + util.logexc(LOG, "Failed to get system UUID", e) + + sys_type_cmd = [dmidecode_path, "-s", "system-product-name"] + try: + LOG.debug("Determining hypervisor product name via dmidecode") + (sys_type, _err) = util.subp(sys_type_cmd) + except Exception as e: + util.logexc(LOG, "Failed to get system UUID", e) + + return sys_uuid.lower(), sys_type + + +# Used to match classes to dependencies +datasources = [ + (DataSourceSmartOS, (sources.DEP_FILESYSTEM, sources.DEP_NETWORK)), +] + + +# Return a list of data sources that match this set of dependencies +def get_datasource_list(depends): + return sources.list_from_depends(depends, datasources) diff --git a/cloudinit/sources/__init__.py b/cloudinit/sources/__init__.py index 0bad4c8b..b0e43954 100644 --- a/cloudinit/sources/__init__.py +++ b/cloudinit/sources/__init__.py @@ -25,6 +25,7 @@ import os from cloudinit import importer from cloudinit import log as logging +from cloudinit import type_utils from cloudinit import user_data as ud from cloudinit import util @@ -52,7 +53,7 @@ class DataSource(object): self.userdata = None self.metadata = None self.userdata_raw = None - name = util.obj_name(self) + name = type_utils.obj_name(self) if name.startswith(DS_PREFIX): name = name[len(DS_PREFIX):] self.ds_cfg = util.get_cfg_by_path(self.sys_cfg, @@ -62,6 +63,9 @@ class DataSource(object): else: self.ud_proc = ud_proc + def __str__(self): + return type_utils.obj_name(self) + def get_userdata(self, apply_filter=False): if self.userdata is None: self.userdata = self.ud_proc.process(self.get_userdata_raw()) @@ -131,7 +135,8 @@ class DataSource(object): @property def availability_zone(self): - return self.metadata.get('availability-zone') + return self.metadata.get('availability-zone', + self.metadata.get('availability_zone')) def get_instance_id(self): if not self.metadata or 'instance-id' not in self.metadata: @@ -221,7 +226,7 @@ def normalize_pubkey_data(pubkey_data): def find_source(sys_cfg, distro, paths, ds_deps, cfg_list, pkg_list): ds_list = list_sources(cfg_list, ds_deps, pkg_list) - ds_names = [util.obj_name(f) for f in ds_list] + ds_names = [type_utils.obj_name(f) for f in ds_list] LOG.debug("Searching for data source in: %s", ds_names) for cls in ds_list: @@ -229,7 +234,7 @@ def find_source(sys_cfg, distro, paths, ds_deps, cfg_list, pkg_list): LOG.debug("Seeing if we can get any data from %s", cls) s = cls(sys_cfg, distro, paths) if s.get_data(): - return (s, util.obj_name(cls)) + return (s, type_utils.obj_name(cls)) except Exception: util.logexc(LOG, "Getting data from %s failed", cls) diff --git a/cloudinit/ssh_util.py b/cloudinit/ssh_util.py index dd6b742f..70a577bc 100644 --- a/cloudinit/ssh_util.py +++ b/cloudinit/ssh_util.py @@ -19,9 +19,6 @@ # You should have received a copy of the GNU General Public License # along with this program. If not, see <http://www.gnu.org/licenses/>. -from StringIO import StringIO - -import csv import os import pwd @@ -33,6 +30,15 @@ LOG = logging.getLogger(__name__) # See: man sshd_config DEF_SSHD_CFG = "/etc/ssh/sshd_config" +# taken from openssh source key.c/key_type_from_name +VALID_KEY_TYPES = ("rsa", "dsa", "ssh-rsa", "ssh-dss", "ecdsa", + "ssh-rsa-cert-v00@openssh.com", "ssh-dss-cert-v00@openssh.com", + "ssh-rsa-cert-v00@openssh.com", "ssh-dss-cert-v00@openssh.com", + "ssh-rsa-cert-v01@openssh.com", "ssh-dss-cert-v01@openssh.com", + "ecdsa-sha2-nistp256-cert-v01@openssh.com", + "ecdsa-sha2-nistp384-cert-v01@openssh.com", + "ecdsa-sha2-nistp521-cert-v01@openssh.com") + class AuthKeyLine(object): def __init__(self, source, keytype=None, base64=None, @@ -43,11 +49,8 @@ class AuthKeyLine(object): self.keytype = keytype self.source = source - def empty(self): - if (not self.base64 and - not self.comment and not self.keytype and not self.options): - return True - return False + def valid(self): + return (self.base64 and self.keytype) def __str__(self): toks = [] @@ -107,62 +110,47 @@ class AuthKeyLineParser(object): i = i + 1 options = ent[0:i] - options_lst = [] - - # Now use a csv parser to pull the options - # out of the above string that we just found an endpoint for. - # - # No quoting so we don't mess up any of the quoting that - # is already there. - reader = csv.reader(StringIO(options), quoting=csv.QUOTE_NONE) - for row in reader: - for e in row: - # Only keep non-empty csv options - e = e.strip() - if e: - options_lst.append(e) - - # Now take the rest of the items before the string - # as long as there is room to do this... - toks = [] - if i + 1 < len(ent): - rest = ent[i + 1:] - toks = rest.split(None, 2) - return (options_lst, toks) - - def _form_components(self, src_line, toks, options=None): - components = {} - if len(toks) == 1: - components['base64'] = toks[0] - elif len(toks) == 2: - components['base64'] = toks[0] - components['comment'] = toks[1] - elif len(toks) == 3: - components['keytype'] = toks[0] - components['base64'] = toks[1] - components['comment'] = toks[2] - components['options'] = options - if not components: - return AuthKeyLine(src_line) - else: - return AuthKeyLine(src_line, **components) - def parse(self, src_line, def_opt=None): + # Return the rest of the string in 'remain' + remain = ent[i:].lstrip() + return (options, remain) + + def parse(self, src_line, options=None): + # modeled after opensshes auth2-pubkey.c:user_key_allowed2 line = src_line.rstrip("\r\n") if line.startswith("#") or line.strip() == '': return AuthKeyLine(src_line) - else: - ent = line.strip() - toks = ent.split(None, 3) - if len(toks) < 4: - return self._form_components(src_line, toks, def_opt) - else: - (options, toks) = self._extract_options(ent) - if options: - options = ",".join(options) - else: - options = def_opt - return self._form_components(src_line, toks, options) + + def parse_ssh_key(ent): + # return ketype, key, [comment] + toks = ent.split(None, 2) + if len(toks) < 2: + raise TypeError("To few fields: %s" % len(toks)) + if toks[0] not in VALID_KEY_TYPES: + raise TypeError("Invalid keytype %s" % toks[0]) + + # valid key type and 2 or 3 fields: + if len(toks) == 2: + # no comment in line + toks.append("") + + return toks + + ent = line.strip() + try: + (keytype, base64, comment) = parse_ssh_key(ent) + except TypeError: + (keyopts, remain) = self._extract_options(ent) + if options is None: + options = keyopts + + try: + (keytype, base64, comment) = parse_ssh_key(remain) + except TypeError: + return AuthKeyLine(src_line) + + return AuthKeyLine(src_line, keytype=keytype, base64=base64, + comment=comment, options=options) def parse_authorized_keys(fname): @@ -186,11 +174,11 @@ def update_authorized_keys(old_entries, keys): for i in range(0, len(old_entries)): ent = old_entries[i] - if ent.empty() or not ent.base64: + if not ent.valid(): continue # Replace those with the same base64 for k in keys: - if k.empty() or not k.base64: + if not ent.valid(): continue if k.base64 == ent.base64: # Replace it with our better one @@ -241,15 +229,13 @@ def extract_authorized_keys(username): except (IOError, OSError): # Give up and use a default key filename auth_key_fn = os.path.join(ssh_dir, 'authorized_keys') - util.logexc(LOG, ("Failed extracting 'AuthorizedKeysFile'" - " in ssh config" - " from %r, using 'AuthorizedKeysFile' file" - " %r instead"), - DEF_SSHD_CFG, auth_key_fn) + util.logexc(LOG, "Failed extracting 'AuthorizedKeysFile' in ssh " + "config from %r, using 'AuthorizedKeysFile' file " + "%r instead", DEF_SSHD_CFG, auth_key_fn) return (auth_key_fn, parse_authorized_keys(auth_key_fn)) -def setup_user_keys(keys, username, key_prefix): +def setup_user_keys(keys, username, options=None): # Make sure the users .ssh dir is setup accordingly (ssh_dir, pwent) = users_ssh_info(username) if not os.path.isdir(ssh_dir): @@ -260,7 +246,7 @@ def setup_user_keys(keys, username, key_prefix): parser = AuthKeyLineParser() key_entries = [] for k in keys: - key_entries.append(parser.parse(str(k), def_opt=key_prefix)) + key_entries.append(parser.parse(str(k), options=options)) # Extract the old and make the new (auth_key_fn, auth_key_entries) = extract_authorized_keys(username) diff --git a/cloudinit/stages.py b/cloudinit/stages.py index d7d1dea0..3e49e8c5 100644 --- a/cloudinit/stages.py +++ b/cloudinit/stages.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2012 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # Copyright (C) 2012 Yahoo! Inc. # # Author: Scott Moser <scott.moser@canonical.com> @@ -43,6 +43,7 @@ from cloudinit import helpers from cloudinit import importer from cloudinit import log as logging from cloudinit import sources +from cloudinit import type_utils from cloudinit import util LOG = logging.getLogger(__name__) @@ -153,9 +154,8 @@ class Init(object): try: util.chownbyname(log_file, u, g) except OSError: - util.logexc(LOG, ("Unable to change the ownership" - " of %s to user %s, group %s"), - log_file, u, g) + util.logexc(LOG, "Unable to change the ownership of %s to " + "user %s, group %s", log_file, u, g) def read_cfg(self, extra_fns=None): # None check so that we don't keep on re-loading if empty @@ -211,7 +211,7 @@ class Init(object): # Any config provided??? pkg_list = self.cfg.get('datasource_pkg_list') or [] # Add the defaults at the end - for n in ['', util.obj_name(sources)]: + for n in ['', type_utils.obj_name(sources)]: if n not in pkg_list: pkg_list.append(n) cfg_list = self.cfg.get('datasource_list') or [] @@ -271,7 +271,7 @@ class Init(object): dp = self.paths.get_cpath('data') # Write what the datasource was and is.. - ds = "%s: %s" % (util.obj_name(self.datasource), self.datasource) + ds = "%s: %s" % (type_utils.obj_name(self.datasource), self.datasource) previous_ds = None ds_fn = os.path.join(idir, 'datasource') try: @@ -344,12 +344,13 @@ class Init(object): cdir = self.paths.get_cpath("handlers") idir = self._get_ipath("handlers") - # Add the path to the plugins dir to the top of our list for import - # instance dir should be read before cloud-dir - if cdir and cdir not in sys.path: - sys.path.insert(0, cdir) - if idir and idir not in sys.path: - sys.path.insert(0, idir) + # Add the path to the plugins dir to the top of our list for importing + # new handlers. + # + # Note(harlowja): instance dir should be read before cloud-dir + for d in [cdir, idir]: + if d and d not in sys.path: + sys.path.insert(0, d) # Ensure datasource fetched before activation (just incase) user_data_msg = self.datasource.get_userdata(True) @@ -357,24 +358,34 @@ class Init(object): # This keeps track of all the active handlers c_handlers = helpers.ContentHandlers() - # Add handlers in cdir - potential_handlers = util.find_modules(cdir) - for (fname, mod_name) in potential_handlers.iteritems(): - try: - mod_locs = importer.find_module(mod_name, [''], - ['list_types', - 'handle_part']) - if not mod_locs: - LOG.warn(("Could not find a valid user-data handler" - " named %s in file %s"), mod_name, fname) - continue - mod = importer.import_module(mod_locs[0]) - mod = handlers.fixup_handler(mod) - types = c_handlers.register(mod) - LOG.debug("Added handler for %s from %s", types, fname) - except: - util.logexc(LOG, "Failed to register handler from %s", fname) - + def register_handlers_in_dir(path): + # Attempts to register any handler modules under the given path. + if not path or not os.path.isdir(path): + return + potential_handlers = util.find_modules(path) + for (fname, mod_name) in potential_handlers.iteritems(): + try: + mod_locs = importer.find_module(mod_name, [''], + ['list_types', + 'handle_part']) + if not mod_locs: + LOG.warn(("Could not find a valid user-data handler" + " named %s in file %s"), mod_name, fname) + continue + mod = importer.import_module(mod_locs[0]) + mod = handlers.fixup_handler(mod) + types = c_handlers.register(mod) + LOG.debug("Added handler for %s from %s", types, fname) + except Exception: + util.logexc(LOG, "Failed to register handler from %s", + fname) + + # Add any handlers in the cloud-dir + register_handlers_in_dir(cdir) + + # Register any other handlers that come from the default set. This + # is done after the cloud-dir handlers so that the cdir modules can + # take over the default user-data handler content-types. def_handlers = self._default_userdata_handlers() applied_def_handlers = c_handlers.register_defaults(def_handlers) if applied_def_handlers: @@ -383,36 +394,51 @@ class Init(object): # Form our cloud interface data = self.cloudify() - # Init the handlers first - called = [] - for (_ctype, mod) in c_handlers.iteritems(): - if mod in called: - continue - handlers.call_begin(mod, data, frequency) - called.append(mod) - - # Walk the user data - part_data = { - 'handlers': c_handlers, - # Any new handlers that are encountered get writen here - 'handlerdir': idir, - 'data': data, - # The default frequency if handlers don't have one - 'frequency': frequency, - # This will be used when new handlers are found - # to help write there contents to files with numbered - # names... - 'handlercount': 0, - } - handlers.walk(user_data_msg, handlers.walker_callback, data=part_data) + def init_handlers(): + # Init the handlers first + for (_ctype, mod) in c_handlers.iteritems(): + if mod in c_handlers.initialized: + # Avoid initing the same module twice (if said module + # is registered to more than one content-type). + continue + handlers.call_begin(mod, data, frequency) + c_handlers.initialized.append(mod) + + def walk_handlers(): + # Walk the user data + part_data = { + 'handlers': c_handlers, + # Any new handlers that are encountered get writen here + 'handlerdir': idir, + 'data': data, + # The default frequency if handlers don't have one + 'frequency': frequency, + # This will be used when new handlers are found + # to help write there contents to files with numbered + # names... + 'handlercount': 0, + } + handlers.walk(user_data_msg, handlers.walker_callback, + data=part_data) + + def finalize_handlers(): + # Give callbacks opportunity to finalize + for (_ctype, mod) in c_handlers.iteritems(): + if mod not in c_handlers.initialized: + # Said module was never inited in the first place, so lets + # not attempt to finalize those that never got called. + continue + c_handlers.initialized.remove(mod) + try: + handlers.call_end(mod, data, frequency) + except: + util.logexc(LOG, "Failed to finalize handler: %s", mod) - # Give callbacks opportunity to finalize - called = [] - for (_ctype, mod) in c_handlers.iteritems(): - if mod in called: - continue - handlers.call_end(mod, data, frequency) - called.append(mod) + try: + init_handlers() + walk_handlers() + finally: + finalize_handlers() # Perform post-consumption adjustments so that # modules that run during the init stage reflect @@ -488,7 +514,7 @@ class Modules(object): else: raise TypeError(("Failed to read '%s' item in config," " unknown type %s") % - (item, util.obj_name(item))) + (item, type_utils.obj_name(item))) return module_list def _fixup_modules(self, raw_mods): @@ -506,7 +532,7 @@ class Modules(object): # Reset it so when ran it will get set to a known value freq = None mod_locs = importer.find_module(mod_name, - ['', util.obj_name(config)], + ['', type_utils.obj_name(config)], ['handle']) if not mod_locs: LOG.warn("Could not find module named %s", mod_name) diff --git a/cloudinit/type_utils.py b/cloudinit/type_utils.py new file mode 100644 index 00000000..2decbfc5 --- /dev/null +++ b/cloudinit/type_utils.py @@ -0,0 +1,34 @@ +# vi: ts=4 expandtab +# +# Copyright (C) 2012 Canonical Ltd. +# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012 Yahoo! Inc. +# +# Author: Scott Moser <scott.moser@canonical.com> +# Author: Juerg Haefliger <juerg.haefliger@hp.com> +# Author: Joshua Harlow <harlowja@yahoo-inc.com> +# +# This program is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License version 3, as +# published by the Free Software Foundation. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see <http://www.gnu.org/licenses/>. +# +# pylint: disable=C0302 + +import types + + +def obj_name(obj): + if isinstance(obj, (types.TypeType, + types.ModuleType, + types.FunctionType, + types.LambdaType)): + return str(obj.__name__) + return obj_name(obj.__class__) diff --git a/cloudinit/url_helper.py b/cloudinit/url_helper.py index f3e3fd7e..19a30409 100644 --- a/cloudinit/url_helper.py +++ b/cloudinit/url_helper.py @@ -20,43 +20,55 @@ # You should have received a copy of the GNU General Public License # along with this program. If not, see <http://www.gnu.org/licenses/>. -from contextlib import closing - -import errno -import socket import time -import urllib -import urllib2 + +import requests +from requests import exceptions + +from urlparse import (urlparse, urlunparse) from cloudinit import log as logging from cloudinit import version LOG = logging.getLogger(__name__) +# Check if requests has ssl support (added in requests >= 0.8.8) +SSL_ENABLED = False +CONFIG_ENABLED = False # This was added in 0.7 (but taken out in >=1.0) +try: + from distutils.version import LooseVersion + import pkg_resources + _REQ = pkg_resources.get_distribution('requests') + _REQ_VER = LooseVersion(_REQ.version) # pylint: disable=E1103 + if _REQ_VER >= LooseVersion('0.8.8'): + SSL_ENABLED = True + if _REQ_VER >= LooseVersion('0.7.0') and _REQ_VER < LooseVersion('1.0.0'): + CONFIG_ENABLED = True +except: + pass + + +def _cleanurl(url): + parsed_url = list(urlparse(url, scheme='http')) # pylint: disable=E1123 + if not parsed_url[1] and parsed_url[2]: + # Swap these since this seems to be a common + # occurrence when given urls like 'www.google.com' + parsed_url[1] = parsed_url[2] + parsed_url[2] = '' + return urlunparse(parsed_url) -class UrlResponse(object): - def __init__(self, status_code, contents=None, headers=None): - self._status_code = status_code - self._contents = contents - self._headers = headers - @property - def code(self): - return self._status_code +class UrlResponse(object): + def __init__(self, response): + self._response = response @property def contents(self): - return self._contents + return self._response.content @property - def headers(self): - return self._headers - - def __str__(self): - if not self.contents: - return '' - else: - return str(self.contents) + def url(self): + return self._response.url def ok(self, redirects_ok=False): upper = 300 @@ -67,72 +79,130 @@ class UrlResponse(object): else: return False + @property + def headers(self): + return self._response.headers -def readurl(url, data=None, timeout=None, - retries=0, sec_between=1, headers=None): - - req_args = {} - req_args['url'] = url - if data is not None: - req_args['data'] = urllib.urlencode(data) + @property + def code(self): + return self._response.status_code + def __str__(self): + return self.contents + + +class UrlError(IOError): + def __init__(self, cause, code=None, headers=None): + IOError.__init__(self, str(cause)) + self.cause = cause + self.code = code + self.headers = headers + if self.headers is None: + self.headers = {} + + +def readurl(url, data=None, timeout=None, retries=0, sec_between=1, + headers=None, headers_cb=None, ssl_details=None, + check_status=True, allow_redirects=True): + url = _cleanurl(url) + req_args = { + 'url': url, + } + scheme = urlparse(url).scheme # pylint: disable=E1101 + if scheme == 'https' and ssl_details: + if not SSL_ENABLED: + LOG.warn("SSL is not enabled, cert. verification can not occur!") + else: + if 'ca_certs' in ssl_details and ssl_details['ca_certs']: + req_args['verify'] = ssl_details['ca_certs'] + else: + req_args['verify'] = True + if 'cert_file' in ssl_details and 'key_file' in ssl_details: + req_args['cert'] = [ssl_details['cert_file'], + ssl_details['key_file']] + elif 'cert_file' in ssl_details: + req_args['cert'] = str(ssl_details['cert_file']) + + req_args['allow_redirects'] = allow_redirects + req_args['method'] = 'GET' + if timeout is not None: + req_args['timeout'] = max(float(timeout), 0) + if data: + req_args['method'] = 'POST' + # It doesn't seem like config + # was added in older library versions (or newer ones either), thus we + # need to manually do the retries if it wasn't... + if CONFIG_ENABLED: + req_config = { + 'store_cookies': False, + } + # Don't use the retry support built-in + # since it doesn't allow for 'sleep_times' + # in between tries.... + # if retries: + # req_config['max_retries'] = max(int(retries), 0) + req_args['config'] = req_config + manual_tries = 1 + if retries: + manual_tries = max(int(retries) + 1, 1) if not headers: headers = { 'User-Agent': 'Cloud-Init/%s' % (version.version_string()), } - - req_args['headers'] = headers - req = urllib2.Request(**req_args) - - retries = max(retries, 0) - attempts = retries + 1 - - excepts = [] - LOG.debug(("Attempting to open '%s' with %s attempts" - " (%s retries, timeout=%s) to be performed"), - url, attempts, retries, timeout) - open_args = {} - if timeout is not None: - open_args['timeout'] = int(timeout) - for i in range(0, attempts): + if not headers_cb: + def _cb(url): + return headers + headers_cb = _cb + + if data: + # Do this after the log (it might be large) + req_args['data'] = data + if sec_between is None: + sec_between = -1 + excps = [] + # Handle retrying ourselves since the built-in support + # doesn't handle sleeping between tries... + for i in range(0, manual_tries): try: - with closing(urllib2.urlopen(req, **open_args)) as rh: - content = rh.read() - status = rh.getcode() - if status is None: - # This seems to happen when files are read... - status = 200 - headers = {} - if rh.headers: - headers = dict(rh.headers) - LOG.debug("Read from %s (%s, %sb) after %s attempts", - url, status, len(content), (i + 1)) - return UrlResponse(status, content, headers) - except urllib2.HTTPError as e: - excepts.append(e) - except urllib2.URLError as e: - # This can be a message string or - # another exception instance - # (socket.error for remote URLs, OSError for local URLs). - if (isinstance(e.reason, (OSError)) and - e.reason.errno == errno.ENOENT): - excepts.append(e.reason) + req_args['headers'] = headers_cb(url) + filtered_req_args = {} + for (k, v) in req_args.items(): + if k == 'data': + continue + filtered_req_args[k] = v + + LOG.debug("[%s/%s] open '%s' with %s configuration", i, + manual_tries, url, filtered_req_args) + + r = requests.request(**req_args) + if check_status: + r.raise_for_status() # pylint: disable=E1103 + LOG.debug("Read from %s (%s, %sb) after %s attempts", url, + r.status_code, len(r.content), # pylint: disable=E1103 + (i + 1)) + # Doesn't seem like we can make it use a different + # subclass for responses, so add our own backward-compat + # attrs + return UrlResponse(r) + except exceptions.RequestException as e: + if (isinstance(e, (exceptions.HTTPError)) + and hasattr(e, 'response') # This appeared in v 0.10.8 + and hasattr(e.response, 'status_code')): + excps.append(UrlError(e, code=e.response.status_code, + headers=e.response.headers)) else: - excepts.append(e) - except Exception as e: - excepts.append(e) - if i + 1 < attempts: - LOG.debug("Please wait %s seconds while we wait to try again", - sec_between) - time.sleep(sec_between) - - # Didn't work out - LOG.debug("Failed reading from %s after %s attempts", url, attempts) - - # It must of errored at least once for code - # to get here so re-raise the last error - LOG.debug("%s errors occured, re-raising the last one", len(excepts)) - raise excepts[-1] + excps.append(UrlError(e)) + if SSL_ENABLED and isinstance(e, exceptions.SSLError): + # ssl exceptions are not going to get fixed by waiting a + # few seconds + break + if i + 1 < manual_tries and sec_between > 0: + LOG.debug("Please wait %s seconds while we wait to try again", + sec_between) + time.sleep(sec_between) + if excps: + raise excps[-1] + return None # Should throw before this... def wait_for_url(urls, max_wait=None, timeout=None, @@ -143,7 +213,7 @@ def wait_for_url(urls, max_wait=None, timeout=None, max_wait: roughly the maximum time to wait before giving up The max time is *actually* len(urls)*timeout as each url will be tried once and given the timeout provided. - timeout: the timeout provided to urllib2.urlopen + timeout: the timeout provided to urlopen status_cb: call method with string message when a url is not available headers_cb: call method with single argument of url to get headers for request. @@ -190,36 +260,40 @@ def wait_for_url(urls, max_wait=None, timeout=None, timeout = int((start_time + max_wait) - now) reason = "" + e = None try: if headers_cb is not None: headers = headers_cb(url) else: headers = {} - resp = readurl(url, headers=headers, timeout=timeout) - if not resp.contents: - reason = "empty response [%s]" % (resp.code) - e = ValueError(reason) - elif not resp.ok(): - reason = "bad status code [%s]" % (resp.code) - e = ValueError(reason) + response = readurl(url, headers=headers, timeout=timeout, + check_status=False) + if not response.contents: + reason = "empty response [%s]" % (response.code) + e = UrlError(ValueError(reason), + code=response.code, headers=response.headers) + elif not response.ok(): + reason = "bad status code [%s]" % (response.code) + e = UrlError(ValueError(reason), + code=response.code, headers=response.headers) else: return url - except urllib2.HTTPError as e: - reason = "http error [%s]" % e.code - except urllib2.URLError as e: - reason = "url error [%s]" % e.reason - except socket.timeout as e: - reason = "socket timeout [%s]" % e + except UrlError as e: + reason = "request error [%s]" % e except Exception as e: reason = "unexpected error [%s]" % e time_taken = int(time.time() - start_time) status_msg = "Calling '%s' failed [%s/%ss]: %s" % (url, - time_taken, - max_wait, reason) + time_taken, + max_wait, + reason) status_cb(status_msg) if exception_cb: + # This can be used to alter the headers that will be sent + # in the future, for example this is what the MAAS datasource + # does. exception_cb(msg=status_msg, exception=e) if timeup(max_wait, start_time): diff --git a/cloudinit/user_data.py b/cloudinit/user_data.py index 58827e3d..d49ea094 100644 --- a/cloudinit/user_data.py +++ b/cloudinit/user_data.py @@ -23,13 +23,14 @@ import os import email + from email.mime.base import MIMEBase from email.mime.multipart import MIMEMultipart +from email.mime.nonmultipart import MIMENonMultipart from email.mime.text import MIMEText from cloudinit import handlers from cloudinit import log as logging -from cloudinit import url_helper from cloudinit import util LOG = logging.getLogger(__name__) @@ -49,6 +50,18 @@ ARCHIVE_TYPES = ["text/cloud-config-archive"] UNDEF_TYPE = "text/plain" ARCHIVE_UNDEF_TYPE = "text/cloud-config" +# This seems to hit most of the gzip possible content types. +DECOMP_TYPES = [ + 'application/gzip', + 'application/gzip-compressed', + 'application/gzipped', + 'application/x-compress', + 'application/x-compressed', + 'application/x-gunzip', + 'application/x-gzip', + 'application/x-gzip-compressed', +] + # Msg header used to track attachments ATTACHMENT_FIELD = 'Number-Attachments' @@ -57,9 +70,21 @@ ATTACHMENT_FIELD = 'Number-Attachments' EXAMINE_FOR_LAUNCH_INDEX = ["text/cloud-config"] +def _replace_header(msg, key, value): + del msg[key] + msg[key] = value + + +def _set_filename(msg, filename): + del msg['Content-Disposition'] + msg.add_header('Content-Disposition', + 'attachment', filename=str(filename)) + + class UserDataProcessor(object): def __init__(self, paths): self.paths = paths + self.ssl_details = util.fetch_ssl_details(paths) def process(self, blob): accumulating_msg = MIMEMultipart() @@ -67,6 +92,10 @@ class UserDataProcessor(object): return accumulating_msg def _process_msg(self, base_msg, append_msg): + + def find_ctype(payload): + return handlers.type_from_starts_with(payload) + for part in base_msg.walk(): if is_skippable(part): continue @@ -74,21 +103,51 @@ class UserDataProcessor(object): ctype = None ctype_orig = part.get_content_type() payload = part.get_payload(decode=True) + was_compressed = False + + # When the message states it is of a gzipped content type ensure + # that we attempt to decode said payload so that the decompressed + # data can be examined (instead of the compressed data). + if ctype_orig in DECOMP_TYPES: + try: + payload = util.decomp_gzip(payload, quiet=False) + # At this point we don't know what the content-type is + # since we just decompressed it. + ctype_orig = None + was_compressed = True + except util.DecompressionError as e: + LOG.warn("Failed decompressing payload from %s of length" + " %s due to: %s", ctype_orig, len(payload), e) + continue + # Attempt to figure out the payloads content-type if not ctype_orig: ctype_orig = UNDEF_TYPE - if ctype_orig in TYPE_NEEDED: - ctype = handlers.type_from_starts_with(payload) - + ctype = find_ctype(payload) if ctype is None: ctype = ctype_orig + # In the case where the data was compressed, we want to make sure + # that we create a new message that contains the found content + # type with the uncompressed content since later traversals of the + # messages will expect a part not compressed. + if was_compressed: + maintype, subtype = ctype.split("/", 1) + n_part = MIMENonMultipart(maintype, subtype) + n_part.set_payload(payload) + # Copy various headers from the old part to the new one, + # but don't include all the headers since some are not useful + # after decoding and decompression. + if part.get_filename(): + _set_filename(n_part, part.get_filename()) + for h in ('Launch-Index',): + if h in part: + _replace_header(n_part, h, str(part[h])) + part = n_part + if ctype != ctype_orig: - if CONTENT_TYPE in part: - part.replace_header(CONTENT_TYPE, ctype) - else: - part[CONTENT_TYPE] = ctype + _replace_header(part, CONTENT_TYPE, ctype) if ctype in INCLUDE_TYPES: self._do_include(payload, append_msg) @@ -98,12 +157,9 @@ class UserDataProcessor(object): self._explode_archive(payload, append_msg) continue - # Should this be happening, shouldn't + # TODO(harlowja): Should this be happening, shouldn't # the part header be modified and not the base? - if CONTENT_TYPE in base_msg: - base_msg.replace_header(CONTENT_TYPE, ctype) - else: - base_msg[CONTENT_TYPE] = ctype + _replace_header(base_msg, CONTENT_TYPE, ctype) self._attach_part(append_msg, part) @@ -138,8 +194,7 @@ class UserDataProcessor(object): def _process_before_attach(self, msg, attached_id): if not msg.get_filename(): - msg.add_header('Content-Disposition', - 'attachment', filename=PART_FN_TPL % (attached_id)) + _set_filename(msg, PART_FN_TPL % (attached_id)) self._attach_launch_index(msg) def _do_include(self, content, append_msg): @@ -173,7 +228,8 @@ class UserDataProcessor(object): if include_once_on and os.path.isfile(include_once_fn): content = util.load_file(include_once_fn) else: - resp = url_helper.readurl(include_url) + resp = util.read_file_or_url(include_url, + ssl_details=self.ssl_details) if include_once_on and resp.ok(): util.write_file(include_once_fn, str(resp), mode=0600) if resp.ok(): @@ -216,13 +272,15 @@ class UserDataProcessor(object): msg.set_payload(content) if 'filename' in ent: - msg.add_header('Content-Disposition', - 'attachment', filename=ent['filename']) + _set_filename(msg, ent['filename']) if 'launch-index' in ent: msg.add_header('Launch-Index', str(ent['launch-index'])) for header in list(ent.keys()): - if header in ('content', 'filename', 'type', 'launch-index'): + if header.lower() in ('content', 'filename', 'type', + 'launch-index', 'content-disposition', + ATTACHMENT_FIELD.lower(), + CONTENT_TYPE.lower()): continue msg.add_header(header, ent[header]) @@ -237,13 +295,13 @@ class UserDataProcessor(object): outer_msg[ATTACHMENT_FIELD] = '0' if new_count is not None: - outer_msg.replace_header(ATTACHMENT_FIELD, str(new_count)) + _replace_header(outer_msg, ATTACHMENT_FIELD, str(new_count)) fetched_count = 0 try: fetched_count = int(outer_msg.get(ATTACHMENT_FIELD)) except (ValueError, TypeError): - outer_msg.replace_header(ATTACHMENT_FIELD, str(fetched_count)) + _replace_header(outer_msg, ATTACHMENT_FIELD, str(fetched_count)) return fetched_count def _attach_part(self, outer_msg, part): @@ -275,10 +333,7 @@ def convert_string(raw_data, headers=None): if "mime-version:" in data[0:4096].lower(): msg = email.message_from_string(data) for (key, val) in headers.iteritems(): - if key in msg: - msg.replace_header(key, val) - else: - msg[key] = val + _replace_header(msg, key, val) else: mtype = headers.get(CONTENT_TYPE, NOT_MULTIPART_TYPE) maintype, subtype = mtype.split("/", 1) diff --git a/cloudinit/util.py b/cloudinit/util.py index 7b1202a2..e1c51f31 100644 --- a/cloudinit/util.py +++ b/cloudinit/util.py @@ -1,7 +1,7 @@ # vi: ts=4 expandtab # # Copyright (C) 2012 Canonical Ltd. -# Copyright (C) 2012 Hewlett-Packard Development Company, L.P. +# Copyright (C) 2012, 2013 Hewlett-Packard Development Company, L.P. # Copyright (C) 2012 Yahoo! Inc. # # Author: Scott Moser <scott.moser@canonical.com> @@ -43,15 +43,16 @@ import subprocess import sys import tempfile import time -import types import urlparse import yaml from cloudinit import importer from cloudinit import log as logging +from cloudinit import mergers from cloudinit import safeyaml -from cloudinit import url_helper as uhelp +from cloudinit import type_utils +from cloudinit import url_helper from cloudinit import version from cloudinit.settings import (CFG_BUILTIN) @@ -70,6 +71,31 @@ FN_ALLOWED = ('_-.()' + string.digits + string.ascii_letters) CONTAINER_TESTS = ['running-in-container', 'lxc-is-container'] +# Made to have same accessors as UrlResponse so that the +# read_file_or_url can return this or that object and the +# 'user' of those objects will not need to know the difference. +class StringResponse(object): + def __init__(self, contents, code=200): + self.code = code + self.headers = {} + self.contents = contents + self.url = None + + def ok(self, *args, **kwargs): # pylint: disable=W0613 + if self.code != 200: + return False + return True + + def __str__(self): + return self.contents + + +class FileResponse(StringResponse): + def __init__(self, path, contents, code=200): + StringResponse.__init__(self, contents, code=code) + self.url = path + + class ProcessExecutionError(IOError): MESSAGE_TMPL = ('%(description)s\n' @@ -193,12 +219,12 @@ def fork_cb(child_cb, *args): child_cb(*args) os._exit(0) # pylint: disable=W0212 except: - logexc(LOG, ("Failed forking and" - " calling callback %s"), obj_name(child_cb)) + logexc(LOG, "Failed forking and calling callback %s", + type_utils.obj_name(child_cb)) os._exit(1) # pylint: disable=W0212 else: LOG.debug("Forked child %s who will run callback %s", - fid, obj_name(child_cb)) + fid, type_utils.obj_name(child_cb)) def is_true(val, addons=None): @@ -381,6 +407,7 @@ def system_info(): 'release': platform.release(), 'python': platform.python_version(), 'uname': platform.uname(), + 'dist': platform.linux_distribution(), } @@ -460,7 +487,7 @@ def redirect_output(outfmt, errfmt, o_out=None, o_err=None): new_fp = open(arg, owith) elif mode == "|": proc = subprocess.Popen(arg, shell=True, stdin=subprocess.PIPE) - new_fp = proc.stdin + new_fp = proc.stdin # pylint: disable=E1101 else: raise TypeError("Invalid type for output format: %s" % outfmt) @@ -482,7 +509,7 @@ def redirect_output(outfmt, errfmt, o_out=None, o_err=None): new_fp = open(arg, owith) elif mode == "|": proc = subprocess.Popen(arg, shell=True, stdin=subprocess.PIPE) - new_fp = proc.stdin + new_fp = proc.stdin # pylint: disable=E1101 else: raise TypeError("Invalid type for error format: %s" % errfmt) @@ -512,38 +539,19 @@ def make_url(scheme, host, port=None, return urlparse.urlunparse(pieces) -def obj_name(obj): - if isinstance(obj, (types.TypeType, - types.ModuleType, - types.FunctionType, - types.LambdaType)): - return str(obj.__name__) - return obj_name(obj.__class__) - - def mergemanydict(srcs, reverse=False): if reverse: srcs = reversed(srcs) - m_cfg = {} - for a_cfg in srcs: - if a_cfg: - m_cfg = mergedict(m_cfg, a_cfg) - return m_cfg - - -def mergedict(src, cand): - """ - Merge values from C{cand} into C{src}. - If C{src} has a key C{cand} will not override. - Nested dictionaries are merged recursively. - """ - if isinstance(src, dict) and isinstance(cand, dict): - for (k, v) in cand.iteritems(): - if k not in src: - src[k] = v - else: - src[k] = mergedict(src[k], v) - return src + merged_cfg = {} + for cfg in srcs: + if cfg: + # Figure out which mergers to apply... + mergers_to_apply = mergers.dict_extract_mergers(cfg) + if not mergers_to_apply: + mergers_to_apply = mergers.default_mergers() + merger = mergers.construct(mergers_to_apply) + merged_cfg = merger.merge(merged_cfg, cfg) + return merged_cfg @contextlib.contextmanager @@ -618,18 +626,64 @@ def read_optional_seed(fill, base="", ext="", timeout=5): fill['user-data'] = ud fill['meta-data'] = md return True - except OSError as e: + except IOError as e: if e.errno == errno.ENOENT: return False raise -def read_file_or_url(url, timeout=5, retries=10, file_retries=0): +def fetch_ssl_details(paths=None): + ssl_details = {} + # Lookup in these locations for ssl key/cert files + ssl_cert_paths = [ + '/var/lib/cloud/data/ssl', + '/var/lib/cloud/instance/data/ssl', + ] + if paths: + ssl_cert_paths.extend([ + os.path.join(paths.get_ipath_cur('data'), 'ssl'), + os.path.join(paths.get_cpath('data'), 'ssl'), + ]) + ssl_cert_paths = uniq_merge(ssl_cert_paths) + ssl_cert_paths = [d for d in ssl_cert_paths if d and os.path.isdir(d)] + cert_file = None + for d in ssl_cert_paths: + if os.path.isfile(os.path.join(d, 'cert.pem')): + cert_file = os.path.join(d, 'cert.pem') + break + key_file = None + for d in ssl_cert_paths: + if os.path.isfile(os.path.join(d, 'key.pem')): + key_file = os.path.join(d, 'key.pem') + break + if cert_file and key_file: + ssl_details['cert_file'] = cert_file + ssl_details['key_file'] = key_file + elif cert_file: + ssl_details['cert_file'] = cert_file + return ssl_details + + +def read_file_or_url(url, timeout=5, retries=10, + headers=None, data=None, sec_between=1, ssl_details=None, + headers_cb=None): + url = url.lstrip() if url.startswith("/"): url = "file://%s" % url - if url.startswith("file://"): - retries = file_retries - return uhelp.readurl(url, timeout=timeout, retries=retries) + if url.lower().startswith("file://"): + if data: + LOG.warn("Unable to post data to file resource %s", url) + file_path = url[len("file://"):] + return FileResponse(file_path, contents=load_file(file_path)) + else: + return url_helper.readurl(url, + timeout=timeout, + retries=retries, + headers=headers, + headers_cb=headers_cb, + data=data, + sec_between=sec_between, + ssl_details=ssl_details) def load_yaml(blob, default=None, allowed=(dict,)): @@ -644,7 +698,7 @@ def load_yaml(blob, default=None, allowed=(dict,)): # Yes this will just be caught, but thats ok for now... raise TypeError(("Yaml load allows %s root types," " but got %s instead") % - (allowed, obj_name(converted))) + (allowed, type_utils.obj_name(converted))) loaded = converted except (yaml.YAMLError, TypeError, ValueError): if len(blob) == 0: @@ -713,7 +767,7 @@ def read_conf_with_confd(cfgfile): if not isinstance(confd, (str, basestring)): raise TypeError(("Config file %s contains 'conf_d' " "with non-string type %s") % - (cfgfile, obj_name(confd))) + (cfgfile, type_utils.obj_name(confd))) else: confd = str(confd).strip() elif os.path.isdir("%s.d" % cfgfile): @@ -724,7 +778,7 @@ def read_conf_with_confd(cfgfile): # Conf.d settings override input configuration confd_cfg = read_conf_d(confd) - return mergedict(confd_cfg, cfg) + return mergemanydict([confd_cfg, cfg]) def read_cc_from_cmdline(cmdline=None): @@ -846,7 +900,7 @@ def get_cmdline_url(names=('cloud-config-url', 'url'), if not url: return (None, None, None) - resp = uhelp.readurl(url) + resp = read_file_or_url(url) if resp.contents.startswith(starts) and resp.ok(): return (key, url, str(resp)) @@ -879,7 +933,7 @@ def is_resolvable(name): for (_fam, _stype, _proto, cname, sockaddr) in result: badresults[iname].append("%s: %s" % (cname, sockaddr[0])) badips.add(sockaddr[0]) - except socket.gaierror: + except (socket.gaierror, socket.error): pass _DNS_REDIRECT_IP = badips if badresults: @@ -892,7 +946,7 @@ def is_resolvable(name): if addr in _DNS_REDIRECT_IP: return False return True - except socket.gaierror: + except (socket.gaierror, socket.error): return False @@ -1428,7 +1482,7 @@ def subp(args, data=None, rcs=None, env=None, capture=True, shell=False, (out, err) = sp.communicate(data) except OSError as e: raise ProcessExecutionError(cmd=args, reason=e) - rc = sp.returncode + rc = sp.returncode # pylint: disable=E1101 if rc not in rcs: raise ProcessExecutionError(stdout=out, stderr=err, exit_code=rc, @@ -1478,11 +1532,19 @@ def shellify(cmdlist, add_header=True): else: raise RuntimeError(("Unable to shellify type %s" " which is not a list or string") - % (obj_name(args))) + % (type_utils.obj_name(args))) LOG.debug("Shellified %s commands.", cmds_made) return content +def strip_prefix_suffix(line, prefix=None, suffix=None): + if prefix and line.startswith(prefix): + line = line[len(prefix):] + if suffix and line.endswith(suffix): + line = line[:-len(suffix)] + return line + + def is_container(): """ Checks to see if this code running in a container of some sort @@ -1537,7 +1599,7 @@ def get_proc_env(pid): fn = os.path.join("/proc/", str(pid), "environ") try: contents = load_file(fn) - toks = contents.split("\0") + toks = contents.split("\x00") for tok in toks: if tok == "": continue @@ -1593,3 +1655,160 @@ def expand_package_list(version_fmt, pkgs): raise RuntimeError("Invalid package type.") return pkglist + + +def parse_mount_info(path, mountinfo_lines, log=LOG): + """Return the mount information for PATH given the lines from + /proc/$$/mountinfo.""" + + path_elements = [e for e in path.split('/') if e] + devpth = None + fs_type = None + match_mount_point = None + match_mount_point_elements = None + for i, line in enumerate(mountinfo_lines): + parts = line.split() + + # Completely fail if there is anything in any line that is + # unexpected, as continuing to parse past a bad line could + # cause an incorrect result to be returned, so it's better + # return nothing than an incorrect result. + + # The minimum number of elements in a valid line is 10. + if len(parts) < 10: + log.debug("Line %d has two few columns (%d): %s", + i + 1, len(parts), line) + return None + + mount_point = parts[4] + mount_point_elements = [e for e in mount_point.split('/') if e] + + # Ignore mounts deeper than the path in question. + if len(mount_point_elements) > len(path_elements): + continue + + # Ignore mounts where the common path is not the same. + l = min(len(mount_point_elements), len(path_elements)) + if mount_point_elements[0:l] != path_elements[0:l]: + continue + + # Ignore mount points higher than an already seen mount + # point. + if (match_mount_point_elements is not None and + len(match_mount_point_elements) > len(mount_point_elements)): + continue + + # Find the '-' which terminates a list of optional columns to + # find the filesystem type and the path to the device. See + # man 5 proc for the format of this file. + try: + i = parts.index('-') + except ValueError: + log.debug("Did not find column named '-' in line %d: %s", + i + 1, line) + return None + + # Get the path to the device. + try: + fs_type = parts[i + 1] + devpth = parts[i + 2] + except IndexError: + log.debug("Too few columns after '-' column in line %d: %s", + i + 1, line) + return None + + match_mount_point = mount_point + match_mount_point_elements = mount_point_elements + + if devpth and fs_type and match_mount_point: + return (devpth, fs_type, match_mount_point) + else: + return None + + +def get_mount_info(path, log=LOG): + # Use /proc/$$/mountinfo to find the device where path is mounted. + # This is done because with a btrfs filesystem using os.stat(path) + # does not return the ID of the device. + # + # Here, / has a device of 18 (decimal). + # + # $ stat / + # File: '/' + # Size: 234 Blocks: 0 IO Block: 4096 directory + # Device: 12h/18d Inode: 256 Links: 1 + # Access: (0755/drwxr-xr-x) Uid: ( 0/ root) Gid: ( 0/ root) + # Access: 2013-01-13 07:31:04.358011255 +0000 + # Modify: 2013-01-13 18:48:25.930011255 +0000 + # Change: 2013-01-13 18:48:25.930011255 +0000 + # Birth: - + # + # Find where / is mounted: + # + # $ mount | grep ' / ' + # /dev/vda1 on / type btrfs (rw,subvol=@,compress=lzo) + # + # And the device ID for /dev/vda1 is not 18: + # + # $ ls -l /dev/vda1 + # brw-rw---- 1 root disk 253, 1 Jan 13 08:29 /dev/vda1 + # + # So use /proc/$$/mountinfo to find the device underlying the + # input path. + mountinfo_path = '/proc/%s/mountinfo' % os.getpid() + lines = load_file(mountinfo_path).splitlines() + return parse_mount_info(path, lines, log) + + +def which(program): + # Return path of program for execution if found in path + def is_exe(fpath): + return os.path.isfile(fpath) and os.access(fpath, os.X_OK) + + _fpath, _ = os.path.split(program) + if _fpath: + if is_exe(program): + return program + else: + for path in os.environ["PATH"].split(os.pathsep): + path = path.strip('"') + exe_file = os.path.join(path, program) + if is_exe(exe_file): + return exe_file + + return None + + +def log_time(logfunc, msg, func, args=None, kwargs=None, get_uptime=False): + if args is None: + args = [] + if kwargs is None: + kwargs = {} + + start = time.time() + + ustart = None + if get_uptime: + try: + ustart = float(uptime()) + except ValueError: + pass + + try: + ret = func(*args, **kwargs) + finally: + delta = time.time() - start + if ustart is not None: + try: + udelta = float(uptime()) - ustart + except ValueError: + udelta = "N/A" + + tmsg = " took %0.3f seconds" % delta + if get_uptime: + tmsg += "(%0.2f)" % udelta + try: + logfunc(msg + tmsg) + except: + pass + return ret diff --git a/cloudinit/version.py b/cloudinit/version.py index 024d5118..4b29a587 100644 --- a/cloudinit/version.py +++ b/cloudinit/version.py @@ -20,7 +20,7 @@ from distutils import version as vr def version(): - return vr.StrictVersion("0.7.2") + return vr.StrictVersion("0.7.3") def version_string(): |