| Age | Commit message (Collapse) | Author |
|---|
|
This change converts the IPv6 netmask from the network_data.json[1]
format to the CIDR style, <IPv6_addr>/<prefix>.
Using an IPv6 address like ffff:ffff:ffff:ffff:: does not work with
NetworkManager, nor networkscripts.
NetworkManager will ignore the route, logging:
ifcfg-rh: ignoring invalid route at \
"::/:: via fd00:fd00:fd00:2::fffe dev $DEV" \
(/etc/sysconfig/network-scripts/route6-$DEV:3): \
Argument for "::/::" is not ADDR/PREFIX format
Similarly if using networkscripts, ip route fail with error:
Error: inet6 prefix is expected rather than \
"fd00:fd00:fd00::/ffff:ffff:ffff:ffff::".
Also a bit of refactoring ...
cloudinit.net.sysconfig.Route.to_string:
* Move a couple of lines around to reduce repeated code.
* if "ADDRESS" not in key -> continute, so that the
code block following it can be de-indented.
cloudinit.net.network_state:
* Refactors the ipv4_mask_to_net_prefix, ipv6_mask_to_net_prefix
removes mask_to_net_prefix methods. Utilize ipaddress library to
do some of the heavy lifting.
LP: #1959148
|
|
Ensure we don't add duplicated nameserver or searchdomains.
This can happen on OpenBSD because of dhcpleased.
|
|
OpenBSD 7.0 comes with a new service called dhcpleased to manage the DHCP
requests.
|
|
Applied Black and isort, fixed any linting issues, updated tox.ini
and CI.
|
|
|
|
If we set a dhcp server side like this:
$ cat /var/tmp/cloud-init/cloud-init-dhcp-f0rie5tm/dhcp.leases
lease {
...
option classless-static-routes 31.169.254.169.254 0.0.0.0,31.169.254.169.254
10.112.143.127,22.10.112.140 0.0.0.0,0 10.112.140.1;
...
}
cloud-init fails to configure the routes via 'ip route add' because to there are
two different routes for 169.254.169.254:
$ ip -4 route add 192.168.1.1/32 via 0.0.0.0 dev eth0
$ ip -4 route add 192.168.1.1/32 via 10.112.140.248 dev eth0
But NetworkManager can handle such scenario successfully as it uses "ip route append".
So change cloud-init to also use "ip route append" to fix the issue:
$ ip -4 route append 192.168.1.1/32 via 0.0.0.0 dev eth0
$ ip -4 route append 192.168.1.1/32 via 10.112.140.248 dev eth0
Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
RHBZ: #2003231
|
|
This attempts to standardize unit test file location under test/unittests/
such that any source file located at cloudinit/path/to/file.py may have a
corresponding unit test file at test/unittests/path/to/test_file.py.
Noteworthy Comments:
====================
Four different duplicate test files existed:
test_{gpg,util,cc_mounts,cc_resolv_conf}.py
Each of these duplicate file pairs has been merged together. This is a
break in git history for these files.
The test suite appears to have a dependency on test order. Changing test
order causes some tests to fail. This should be rectified, but for now
some tests have been modified in
tests/unittests/config/test_set_passwords.py.
A helper class name starts with "Test" which causes pytest to try
executing it as a test case, which then throws warnings "due to Class
having __init__()". Silence by changing the name of the class.
# helpers.py is imported in many test files, import paths change
cloudinit/tests/helpers.py -> tests/unittests/helpers.py
# Move directories:
cloudinit/distros/tests -> tests/unittests/distros
cloudinit/cmd/devel/tests -> tests/unittests/cmd/devel
cloudinit/cmd/tests -> tests/unittests/cmd/
cloudinit/sources/helpers/tests -> tests/unittests/sources/helpers
cloudinit/sources/tests -> tests/unittests/sources
cloudinit/net/tests -> tests/unittests/net
cloudinit/config/tests -> tests/unittests/config
cloudinit/analyze/tests/ -> tests/unittests/analyze/
# Standardize tests already in tests/unittests/
test_datasource -> sources
test_distros -> distros
test_vmware -> sources/vmware
test_handler -> config # this contains cloudconfig module tests
test_runs -> runs
|
|
Given that there are additional network management tools that we haven't
yet supported with activators, we should log a warning and continue
without network activation here, especially since this was a no-op for
years.
LP: #1948681
|
|
This patch address an issue where the use of the "set-name"
directive caused the networkd renderer to fail.
LP: #1949407
|
|
This patch addresses an issue caused when the v2 network config
directive "set-name" was used in conjunction with interface-
specific DNS settings. The patch adds a test to validate the fix.
For more information please see bug 1946493 as well as the issue
https://github.com/kubernetes-sigs/image-builder/issues/712.
LP: #1946493
|
|
Fixes the following pylint error:
cloudinit/net/renderer.py:12: [E0611(no-name-in-module), ]
No name 'generate_udev_rule' in module 'udev'
Likely a false positive, but we don't really need to keep the imports
relative, so let's convert them to absolute as a workaround.
|
|
Add connectivity_url to Oracle's EphemeralDHCPv4
On bionic, when trying to bring up the EphemeralDHCPv4, it's possible
that we already have a route defined, which will result in an error when
trying to add the DHCP route. Use the connectivity_url to check if we
can reach the metadata service, and if so, skip the EphemeralDHCPv4.
The has_url_connectivity function has also been modified to take
a dict of kwargs to send to readurl.
LP: #1939603
|
|
openEuler Homepage: https://www.openeuler.org/en/
|
|
https://www.cloudlinux.com/
|
|
Also fix search path in networkd
|
|
Add tests for cc_resolv_conf handler
|
|
|
|
This patch updates some indentation in a comment that prevented an
attempt to run the Black formatter (https://github.com/psf/black)
against the cloud-init codebase:
$ find cloudinit -name '*.py' -type f | xargs black -l 79 --check
...
Oh no! 💥 💔 💥
262 files would be reformatted, 19 files would be left unchanged, 1 file would fail to reformat.
The one file that fails to format is cloudinit/net/__init__.py.
With this fix in place, the black command can successfully parse the
file into AST and back again:
$ black -l 79 --check cloudinit/net/__init__.py
would reformat cloudinit/net/__init__.py
Oh no! 💥 💔 💥
1 file would be reformatted.
Normally this patch would be part of such an overall effort, but since
this is the only location that interrupted running the black command,
this author felt it was worth addressing this discrepancy sooner than
later in the case there is subsequent desire to use a standard format
tool such as black.
|
|
|
|
|
|
Virtuozzo Linux is a distro based off of CentOS 8, similar to Alma Linux and Rocky Linux.
|
|
Adds a udev script which will invoke a hotplug hook script on all net
add events. The script will write some udev arguments to a systemd FIFO
socket (to ensure we have only instance of cloud-init running at a
time), which is then read by a new service that calls a new 'cloud-init
devel hotplug-hook' command to handle the new event.
This hotplug-hook command will:
- Fetch the pickled datsource
- Verify that the hotplug event is supported/enabled
- Update the metadata for the datasource
- Ensure the hotplugged device exists within the datasource
- Apply the config change on the datasource metadata
- Bring up the new interface (or apply global network configuration)
- Save the updated metadata back to the pickle cache
Also scattered in some unrelated typing where helpful
|
|
We read the MTU from the subnet entries. With the v1 format, the MTU can
be set at the root level of the interface entry in the `config` section.
Limitation, we won't set the MTU if the interface use DHCP. This
would require a bit of refactoring.
Also simplify/clarify how we pass the target variable in `cloudinit.net.bsd`.
See: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=256309
Reported-by: Andrey Fesenko
|
|
Currently _bring_up_interfaces() is a no-op for any distro using
renderers. We need to be able to support bringing up a single
interfaces, a list of interfaces, and all interfaces. This should be
independent of the renderers, as the network config is often
generated independent of the mechanism used to apply it.
Additionally, I included a refactor to remove
"_supported_write_network_config". We had a confusing call chain of
apply_network_config->_write_network_config->_supported_write_network_config.
The last two have been combined.
|
|
Minor fixes in networkd renderer & fixed corresponding tests
Removed datasource_list for Photon from cloud.cfg.tmpl & added a comment
in cloud.cfg.tmpl about not to use multiline array for datasource_list.
Signed-off-by: Shreenidhi Shedi <sshedi@vmware.com>
|
|
Also added a new (currently experimental) systemd-networkd renderer,
and includes a small refactor to cc_resolv_conf.py to support the
resolved.conf used by systemd-resolved.
|
|
v1 network config currently has no concept of interface-specific DNS,
which is required for certain renderers. To fix this, added an
optional 'interface' key on the v1 nameserver definition. If
specified, it makes the DNS settings specific to the interface.
Otherwise, it will be defined as global DNS as it always has.
Additionally, DNS for v2 wasn't being recognized correctly. For DNS
defined on a particular interface, these settings now also go into the
global DNS settings as they were intended.
|
|
- Mostly based on FreeBSD, the main exception is that
`find_devs_with_on_freebsd` does not work.
- Since we cannot get the CDROM or the partition labels,
`find_devs_with_on_dragonflybsd()` has a more naive approach and
returns all the block devices.
|
|
dhclient output that contains brackets for pxe variables will break
the dhclient parsing regex line. This fix retains the current
functionality while fixing this particular issue.
|
|
Ensure we've got a clean environment before we restart the network.
In some cases, the `sh /etc/netstart` is not enough to restart the
network. A previous default route remains in the route table and
as a result the network is broken.
Also `sh /netstart` does not kill `dhclient`.
The problen happens for instance with OVH OpenStack SBG3.
|
|
Rocky Linux is a RHEL-compatible distribution so all changes that have
been made should be trivial.
|
|
In the case of a static network, we now set the MTU according to the
meta-data.
|
|
AlmaLinux OS is RHEL-compatible so all the changes needed are trivial.
|
|
Update sysconfig configuration to use BONDING_MODULES_OPTS instead of
BONDING_OPTS when on a SUSE system. The sysconfig support requires use
of BONDING_MODULE_OPTS whereas the initscript support that rhel uses
requires BONDING_OPTS.
|
|
When bringing up DHCP-provided static routes, we check for "0.0.0.0/0"
to indicate an unspecified gateway. However, when parsing the static
route in `parse_static_routes`, the gateway is never specified with
a net length, so the "/0" will never happen.
This change updates the gateway check to check only for "0.0.0.0".
|
|
klibc initramfs in debian allows the 'iscsi_target_ip=' cmdline
parameter to specify an iscsi device attachment. This can
cause cloud-init to mis-detect the cmdline paramter as a
networking config.
LP: #1919188
|
|
`get_interfaces` is used to in two ways, broadly: firstly, to determine
the available interfaces when converting cloud network configuration
formats to cloud-init's network configuration formats; and, secondly, to
ensure that any interfaces which are specified in network configuration
are (a) available, and (b) named correctly. The first of these is
unaffected by this commit, as no clouds support Open vSwitch
configuration in their network configuration formats.
For the second, we check that MAC addresses of physical devices are
unique. In some OVS configurations, there are OVS-created devices which
have duplicate MAC addresses, either with each other or with physical
devices. As these interfaces are created by OVS, we can be confident
that (a) they will be available when appropriate, and (b) that OVS will
name them correctly. As such, this commit excludes any OVS-internal
interfaces from the set of interfaces returned by `get_interfaces`.
LP: #1912844
|
|
Route '-net' parameter is incompatible with /32 IPv4 addresses so we
have to use '-host' in that case.
|
|
IPV6_AUTOCONF needs to be set to 'no' on RHEL so NetworkManager can
properly acquire ipv6 address.
rhbz: #1859695
Signed-off-by: Eduardo Otubo <otubo@redhat.com>
|
|
BOOTPROTO needs to be set to 'dhcp' on RHEL so NetworkManager can
properly acquire ipv6 address.
rhbz: #1859695
Signed-off-by: Eduardo Otubo <otubo@redhat.com>
Co-authored-by: Daniel Watkins <oddbloke@ubuntu.com>
Co-authored-by: Scott Moser <smoser@brickies.net>
|
|
Add code so that specifying "wakeonlan: true" actually results in relevant
configuration entry appearing in /etc/network/interfaces, Netplan, and
sysconfig for RHEL and OpenSuse.
Add testcases for the above.
|
|
The static and static6 subnet types for network_data.json were
being ignored by the Openstack handler, this would cause the code to
break and not function properly.
As of today, if a static6 configuration is chosen, the interface will
still eventually be available to receive router advertisements or be set
from NetworkManager to wait for them and cycle the interface in negative
case.
It is safe to assume that if the interface is manually configured to use
static ipv6 address, there's no need to wait for router advertisements.
This patch will set automatically IPV6_AUTOCONF and IPV6_FORCE_ACCEPT_RA
both to "no" in this case.
This patch fixes the specific behavior only for RHEL flavor and
sysconfig renderer. It also introduces new unit tests for the specific
case as well as adjusts some existent tests to be compatible with the
new options. This patch also addresses this problem by assigning the
appropriate subnet type for each case on the openstack handler.
rhbz: #1889635
rhbz: #1889635
Signed-off-by: Eduardo Otubo otubo@redhat.com
|
|
If an OVS bridge was used as the only/primary interface, the 'init'
stage failed with a "Not all expected physical devices present" error,
leaving the system with a broken SSH setup.
LP: #1898997
|
|
enumeration of physical network devices (#591)
|
|
Prior to this change, vlans were rendered in sysconfig with
'TYPE=Ethernet', and incorrectly rendered the PHYSDEV based on
the name of the vlan device rather than the 'link' provided
in the network config.
The change here fixes:
* rendering of TYPE=Ethernet for a vlan
* adds a warning if the configured device name is not supported
per the RHEL 7 docs "11.5. Naming Scheme for VLAN Interfaces"
LP: #1788915
LP: #1826608
RHBZ: #1861871
|
|
Changes:
tox: bump the pylint version to 2.6.0 in the default run
Fix pylint 2.6.0 W0707 warnings (raise-missing-from)
|
|
* DHCP sandboxing failing on noexec mounted /var/tmp
If /var/tmp is mounted with noexec option the DHCP sandboxing will fail
with Permission Denied. This patch simply avoids this error by checking
the exec permission updating the dhcp path in negative case.
rhbz: https://bugzilla.redhat.com/show_bug.cgi?id=1857309
Signed-off-by: Eduardo Otubo <otubo@redhat.com>
* Replacing with os.* calls
* Adding test and removing isfile() useless call.
Co-authored-by: Rick Harding <rharding@mitechie.com>
|
|
* cloudinit: remove global disable of pylint W0107 and fix errors
This includes removing a test class which contained no tests but wasn't
detected as empty because of an errant pass statement.
* .pylintrc: update disable comment to match arguments
|
|
* Refactor `cloudinit.net.wait_for_physdevs` to `cloudinit.distros.networking.Networking.wait_for_physdevs`
* Split the Linux-specific `udevadm_settle` call out to a separate abstract `Networking.settle` method; implement it on `LinuxNetworking` and add a `NotImplementedError` implementation to `BSDNetworking`
* Modify `wait_for_physdevs`s one callsite to use the new location
LP: #1884626
|
|
This includes a fix to a test that had a string concatenation issue, and
so was only testing a prefix of what was intended.
|
