| Age | Commit message (Collapse) | Author |
|---|
|
Update documentation for network-config-format-v2 to reflect the fact that
routes are a per-interface property, not a global network property.
|
|
CA Cert tests will fail on systems that don't have ca-certificates
installed and configured.
Signed-off-by: Daniel Watkins <oddbloke@ubuntu.com>
Signed-off-by: Eduardo Otubo <otubo@redhat.com>
|
|
Add an integration test that roughly mimics many of the manual cloud
SRU tests. Also refactored some of the image setup code to make it
easier to use in non-fixture code.
|
|
Test that we can add optional vendor-data to the seedfrom file in a
NoCloud environment.
Also added the option to pass raise_on_cloudinit_failure through
an instance restart so we get automatic failure checking when
we need to manually reboot.
|
|
For SRU test development, every single time we start a new test run,
we need to first install the PROPOSED version and create an image
snapshot. Instead of automatically deleting a snapshot, add an
integration setting to allow us to keep the snapshot. The end of the
test run will log the image name which can then be used as the
OS_IMAGE in subsequent test runs.
|
|
If a non-default AuthorizedKeysFile is specified in
/etc/ssh/sshd_config, ensure we can still ssh as expected
|
|
This introduces an optional, more complex OS_IMAGE format (`<image
id>::<os>::<release>`) which allows the specification of the OS/OS
release which the given image ID corresponds to. This information is
used to skip tests which do not apply to the image.
This commit is comprised of the following discrete changes:
* introduce the IntegrationImage class, to handle parsing and storing
the new OS_IMAGE format
* support inferring the OS and OS release of Ubuntu series, so that we
can continue to set OS_IMAGE to just a series name and have test
skipping work
* add documentation on Image Selection to integration_tests.rst
* introduce the actual skipping behaviour based on OS marks
* apply the `ubuntu` mark to all tests that should be skipped on
non-Ubuntu operating systems
|
|
This wraps pycloudlib's `BaseInstance.restart` and `BaseInstance.wait`
to pass the same parameters as on launch, to avoid cloud-init failures
on the _reboot_ raising an exception.
|
|
|
|
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>
|
|
Integration tests have been leaving behind snapshot images, so now we
clean them up. Also, in testing, found that in Azure, deleting a
resource group will automatically delete the instance, so if
KEEP_INSTANCE is True, we no longer delete the resource group.
Co-authored-by: Daniel Watkins <oddbloke@ubuntu.com>
|
|
|
|
pycloudlib will stop running commands as root
by default on LXD. To align with that change
and make the behavior consistent with other
clouds we support, our LXD instances will now
run the commands with sudo by default.
|
|
|
|
|
|
During teardown of every cloud instance, run 'cloud-init collect-logs',
then transfer and unpack locally. Two new integration settings have
been added to specify when to perform this action (ALWAYS,
ON_ERROR, NEVER), and where to store these logs.
|
|
Bump the version in cloudinit/version.py to 20.4 and
update ChangeLog.
LP: #1905440
|
|
Xenial tox 2.3.1 still suffers from
https://github.com/tox-dev/tox/issues/208.
As a result we cannot use testenv substvars if we want to support running
tox on xenial systems.
Drop the {[testenv:integration-tests]*} substitutions to avoid tracebacks
when running `tox -e xenial-dev` which has the signature:
"No support for the %s substitution type" % sub_type)
tox.ConfigError: ConfigError: No support for the posargs substitution
type
|
|
Tests previously assumed that when executing commands and transferring
files that user will have root permissions. This change updated
integration testing infrastructure so that is true.
|
|
|
|
See #585
|
|
`IMAGE_SOURCE` was renamed to `CLOUD_INIT_SOURCE` before the testing
framework ever landed, but I did not mirror that change in .travis.yml
before it landed. This addresses that.
|
|
Adds the ability to run the Azure preprovisioned VMs as NIC-less and
then hot-attach them when assigned for reprovision.
The NIC on the preprovisioned VM is hot-detached as soon as it reports
ready and goes into wait for one or more interfaces to be hot-attached.
Once they are attached, cloud-init gets the expected number of NICs (in
case there are more than one) that will be attached from IMDS and waits
until all of them are attached. After all the NICs are attached,
reprovision proceeds as usual.
|
|
Existing config writes keys to /etc/ssh after deleting files matching
a glob that includes certificate files. Since sshd looks for
certificates in the same directory as the keys, a host certificate
must be placed in this directory. This update enables the certificate's
contents to be specified along with the keys.
Co-authored-by: jonathan lung <lungj@heresjono.com>
Co-authored-by: jonathan lung <jlung@kepler.space>
|
|
As the first test of this SRU cycle, this also introduces the
sru_2020_11 mark to allow us to easily identify the set of tests
generated for this SRU.
|
|
On FreeBSD, if a UFS has trim: (-t) or MAC multilabel: (-l) flag, resize
FS fail, because the _can_skip_ufs_resize check gets tripped up by the
missing options.
This was reported at FreeBSD Bugzilla:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=250496 and as
LP: #1901958
Rather than fixing the parser as in the patches proposed there (and
attempted in #636) this pull-request rips out all of it, and simplifies
the code. We now use `growfs -N` and check if that returns an error. If
it returns the correct kind of error, we can skip the resize, because we
either are at the correct size, or the filesystem in question is broken
or not UFS. If it returns the wrong kind of error, we just re-raise it.
LP: #1901958
|
|
Pushing dmesg log to KVP to help troubleshoot VM boot issues
|
|
IMAGE_SOURCE = 'IN_PLACE' wasn't working previously. Replaced
LXD launch with an init, then mount, then start.
|
|
|
|
cc_set_password will only update the password for the default user if
cfg['password'] is set. The existing code of datasource Azure will fail
to update the default user's password because it does not set that
metadata. If the default user doesn't exist in the image, the current
code works fine because the password is set during user create and
not in cc_set_password
|
|
`-k` will select tests or marks whose names match. `-m` selects only
tests which have the specified mark, so is more appropriate.
|
|
Increase Azure Endpoint HTTP retries to handle
occasional platform network blips.
Introduce a common method http_with_retries
in the azure.py helper, which will serve as
the common HTTP request handler for
all HTTP requests with the Azure endpoint.
This method has builtin retries and
reporting diagnostics logic.
|
|
On systems where the Azure datasource
is a viable platform for crawling metadata,
cloud-init occasionally encounters fatal
irrecoverable errors during the crawling
of the Azure datasource.
When this happens, cloud-init crashes,
and Azure VM provisioning would fail.
However, instead of failing immediately,
the user will continue seeing provisioning
for a long time until it times out with
"OS Provisioning Timed Out" message.
In these situations, cloud-init should
report failure to the Azure datasource
endpoint indicating provisioning failure.
The user will immediately see provisioning
terminate, giving them a much better
failure experience instead of pointlessly
waiting for OS provisioning timeout.
|
|
|
|
This introduces the "ci" mark, used to indicate a test which should run
as part of our CI integration testing run and the integration-tests-ci
tox environment, which runs only those tests. Travis has been adjusted
to use this tox environment.
(All current module tests have been marked with the "ci" mark, but the
one bug test that we have has not.)
|
|
Allow root user to validate the userdata provided to the launched
machine using `cloud-init devel schema --system`
|
|
|
|
This commit does the following:
* introduces the `cloudinit.persistence` module, containing
`CloudInitPickleMixin` which provides lightweight versioning of
objects' pickled representations (and associated testing)
* introduces a basic upgrade testing framework (in
`cloudinit.tests.test_upgrade`) which unpickles pickles from previous
versions of cloud-init (stored in `tests/data/old_pickles`) and tests
invariants that the current cloud-init codebase expects
* uses the versioning framework to address an upgrade issue where
`Distro.networking` could get into an unexpected state, and uses the
upgrade testing framework to confirm that the issue is addressed
|
|
Make sure that gpg works even if the instance has no /dev/tty. This has
been observed on Debian.
LP: #1813396
|
|
|
|
This is a new builtin variable that appeared in Ubuntu in
5.1~rc2-1ubuntu1 and started causing daily build failures.
|
|
|
|
The cloud_tests and the integration_tests (via pycloudlib) have
conflicting requirements. This commit splits up their requirements
files, so we can accurately express the requirements for each.
|
|
|
|
Largely speaking, ds-identify protects from this scenario being
hit, but if DataSourceRbxCloud ran and there was no metadata
disks found (LABEL=CLOUDMD), then it would stacktrace.
The fix is just to clean up the get_md function a little
bit, and the explicitly check for False as a return value in _get_data.
|
|
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.
|
|
We had some inconsistencies between `distro.net.<func>` and
`distro.networking.<func>`. This commit unifies all such instances to
`distro.networking.<func>`.
|
|
FreeBSD lets us read out kernel parameters with kenv(1), a user-space
utility that's shipped in "base" We can use it in place of dmidecode(8),
thus removing the dependency on sysutils/dmidecode, and the restrictions
to i386 and x86_64 architectures that this utility imposes on FreeBSD.
Co-authored-by: Scott Moser <smoser@brickies.net>
|
|
Add a script in travis to output a dot every 10 minute during
and integration test run to prevent travis timeout.
Co-authored-by: Rick Harding <rharding@mitechie.com>
|
|
This allows the cloud-init log to be pushed multiple times during boot,
with the latest lines being pushed each time.
|
