From 0964b42e5117cce640a8ba9102a76fa54a698898 Mon Sep 17 00:00:00 2001 From: Scott Moser Date: Mon, 21 Mar 2016 21:47:24 -0400 Subject: quickly check to see if the previous instance id is still valid This adds a check in cloud-init to see if the existing (cached) datasource is still valid. It relies on support from the Datasource to implement 'check_instance_id'. That method should quickly determine (if possible) if the instance id found in the datasource is still valid. This means that we can still notice new instance ids without depending on a network datasource on every boot. I've also implemented check_instance_id for the superclass and for 3 classes: DataSourceAzure (check dmi data) DataSourceOpenstack (check dmi data) DataSourceNocloud (check the seeded data or kernel command line) LP: #1553815 --- cloudinit/stages.py | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-) (limited to 'cloudinit/stages.py') diff --git a/cloudinit/stages.py b/cloudinit/stages.py index dbcf3d55..edad6450 100644 --- a/cloudinit/stages.py +++ b/cloudinit/stages.py @@ -140,7 +140,7 @@ class Init(object): ] return initial_dirs - def purge_cache(self, rm_instance_lnk=True): + def purge_cache(self, rm_instance_lnk=False): rm_list = [] rm_list.append(self.paths.boot_finished) if rm_instance_lnk: @@ -238,21 +238,29 @@ class Init(object): cfg_list = self.cfg.get('datasource_list') or [] return (cfg_list, pkg_list) - def _get_data_source(self): + def _get_data_source(self, existing): if self.datasource is not NULL_DATA_SOURCE: return self.datasource with events.ReportEventStack( name="check-cache", - description="attempting to read from cache", + description="attempting to read from cache [%s]" % existing, parent=self.reporter) as myrep: ds = self._restore_from_cache() - if ds: - LOG.debug("Restored from cache, datasource: %s", ds) - myrep.description = "restored from cache" + if ds and existing == "trust": + myrep.description = "restored from cache: %s" % ds + elif ds and existing == "check": + if hasattr(ds, 'check_instance_id') and ds.check_instance_id(): + myrep.description = "restored from checked cache: %s" % ds + else: + myrep.description = "cache invalid in datasource: %s" % ds + ds = None else: myrep.description = "no cache found" + LOG.debug(myrep.description) + if not ds: + util.del_file(self.paths.instance_link) (cfg_list, pkg_list) = self._get_datasources() # Deep copy so that user-data handlers can not modify # (which will affect user-data handlers down the line...) @@ -332,8 +340,8 @@ class Init(object): self._reset() return iid - def fetch(self): - return self._get_data_source() + def fetch(self, existing="check"): + return self._get_data_source(existing=existing) def instancify(self): return self._reflect_cur_instance() -- cgit v1.2.3 From 6ce134c1868478345471ba9166f1523f7d9bf19d Mon Sep 17 00:00:00 2001 From: Scott Moser Date: Tue, 22 Mar 2016 03:02:31 -0400 Subject: move some of the pickle loading out of Init, into private methods I plan to re-use these methods later. They stand alone even if they dont end up getting used, though. --- cloudinit/stages.py | 65 ++++++++++++++++++++++++++++------------------------- 1 file changed, 35 insertions(+), 30 deletions(-) (limited to 'cloudinit/stages.py') diff --git a/cloudinit/stages.py b/cloudinit/stages.py index edad6450..c230ec0d 100644 --- a/cloudinit/stages.py +++ b/cloudinit/stages.py @@ -193,40 +193,12 @@ class Init(object): # We try to restore from a current link and static path # by using the instance link, if purge_cache was called # the file wont exist. - pickled_fn = self.paths.get_ipath_cur('obj_pkl') - pickle_contents = None - try: - pickle_contents = util.load_file(pickled_fn, decode=False) - except Exception as e: - if os.path.isfile(pickled_fn): - LOG.warn("failed loading pickle in %s: %s" % (pickled_fn, e)) - pass - - # This is expected so just return nothing - # successfully loaded... - if not pickle_contents: - return None - try: - return pickle.loads(pickle_contents) - except Exception: - util.logexc(LOG, "Failed loading pickled blob from %s", pickled_fn) - return None + return _pkl_load(self.paths.get_ipath_cur('obj_pkl')) def _write_to_cache(self): if self.datasource is NULL_DATA_SOURCE: return False - pickled_fn = self.paths.get_ipath_cur("obj_pkl") - try: - pk_contents = pickle.dumps(self.datasource) - except Exception: - util.logexc(LOG, "Failed pickling datasource %s", self.datasource) - return False - try: - util.write_file(pickled_fn, pk_contents, omode="wb", mode=0o400) - except Exception: - util.logexc(LOG, "Failed pickling datasource to %s", pickled_fn) - return False - return True + return _pkl_store(self.datasource, self.paths.get_ipath_cur("obj_pkl")) def _get_datasources(self): # Any config provided??? @@ -796,3 +768,36 @@ def fetch_base_config(): base_cfgs.append(default_cfg) return util.mergemanydict(base_cfgs) + + +def _pkl_store(obj, fname): + try: + pk_contents = pickle.dumps(obj) + except Exception: + util.logexc(LOG, "Failed pickling datasource %s", obj) + return False + try: + util.write_file(fname, pk_contents, omode="wb", mode=0o400) + except Exception: + util.logexc(LOG, "Failed pickling datasource to %s", fname) + return False + return True + + +def _pkl_load(fname): + pickle_contents = None + try: + pickle_contents = util.load_file(fname, decode=False) + except Exception as e: + if os.path.isfile(fname): + LOG.warn("failed loading pickle in %s: %s" % (fname, e)) + pass + + # This is allowed so just return nothing successfully loaded... + if not pickle_contents: + return None + try: + return pickle.loads(pickle_contents) + except Exception: + util.logexc(LOG, "Failed loading pickled blob from %s", fname) + return None -- cgit v1.2.3