Tripleo implementation mounts /var/log/containers/<name> directories to /var/log/<name> in containers.
Such paths should be dropped, otherwise such logs are collected multiple times for each containers.
When debug is enabled, logs (especially libvirt related logs) become too large to be collected in INFO_DIR.
This patch drops paths mounted from the host.
Change-Id: Ic6cd058671a5cedf8e94200937ea2bc852caf79c
This adjust the regex to capture any other related files
(esp tripleo-container-image-prepare-compute.log). This aligns
with the value added in [1].
[1] https://review.opendev.org/c/openstack/tripleo-ci/+/867966
Change-Id: I5f02bb542e30ffe49ac4e5ea0499dee4f03eafc7
Setuptools v54.1.0 introduces a warning that the use of dash-separated
options in 'setup.cfg' will not be supported in a future version [1].
Get ahead of the issue by replacing the dashes with underscores. Without
this, we see 'UserWarning' messages like the following on new enough
versions of setuptools:
UserWarning: Usage of dash-separated 'description-file' will not be
supported in future versions. Please use the underscore name
'description_file' instead
[1] https://github.com/pypa/setuptools/commit/a2e9ae4cb
Change-Id: I54d4f816d8de765035aa1607740345c13f9b6cc9
In some environment the ssh client is waiting for 10 seconds to
connect to the server, even though the use of an IP.
If we remove the dns from the nsswitch.conf the delay disapairs.
Let's check the config file, to ensure that there is nothing weird
there.
Change-Id: Ia0c2f9548d57bd4ed5091524687795aa3f5c1322
* Added infrared_plugin/main.yml to exclude path as ansible-lint fails as
infrared_plugin/main.yml:5:3: syntax-check[specific]: The field 'hosts'
has an invalid value, which includes an undefined variable. The error
was: 'other' is undefined. 'other' is undefined
* Bump version on ansible-lint
* Set upper limit for tox as version 4 is not well compatible yet
Change-Id: Id60cb45a96a8080c4257d33a0739d0972978045d
New images are using small /tmp volumes
by default which can lead to too many logs
created in a long running job.
Since those jobs does not using sanitizing
the gzip compression can be moved to an
earlier stage relatively easly
which is sufficient to solve the issue.
Closes-Bug: 2013321
Change-Id: Idd31cbf2adf8a8600f556e9b9375fd6f682247b6
Poetry got a new linter rule for pipfile,
which was not followed by isort,
isort stopped supporting old python, but
still released a fix for the stable branch for newer.
The black checker also show some issue,
which is not visible with current fedora37,
Just disabling it for now.
Change-Id: I42b807b03ee79e3f3c8317dfc840cb18ff344e4b
In an effort for ci-logging-processing process the testrepository to
have tempest results on opensearch, the testrepository can not be
gzipped, since logging processing tool doesn't support gzipped files
yet, and also be in the root of the logs directory.
Depends-On: https://review.opendev.org/c/openstack/tripleo-ci/+/859526
Change-Id: Iebba70821f7aec02dc97477f49148064da9aa743
The ansible sanity job was failing because the new node
that runs the job was updated and no longer have python 3.8.
In order to use python 3.10 you need to update ansible-core
to at least 2.13, doing so, it also requires that we add
the ignore-2.13.txt, and update the zuul.d/layout.yaml.
The pylint complains then about use-a-generator, so updating
sova to fix the issue, once the PEP0289 [1] is from 2003,
it should work on all python versions.
The linters job started to fails because gitlab requires now
authentication to clone the flake8 repo, so moving to flake8
repo on github.
1 - https://peps.python.org/pep-0289/
Change-Id: I1370d41b0c4e057c1c65cfd75ff945d648696771
the netstat command overlooked some of the connections/sockets,
preventing efficient debugging when times come to look at connection
states.
Change-Id: I4fb7650568021d89876cbc05a858a29bc2ce4dbe
The "-o cat" removes the timestamp, making it pretty useless when we
want to investigate dropped packets based on time.
Change-Id: I1534a39e01ae9ade52ef0c8cd353d49ab26b68f8
For newer deployments, config download stores artifacts under
~/overcloud-deploy. Those also include mult-cell artifacts.
Collect these directories, and generated static ansible inventories for
it.
Change-Id: Ibb9b5e3a153c64d44ae78b304c372287d0f4cca1
Signed-off-by: Bogdan Dobrelya <bdobreli@redhat.com>
With the current[1] configuration, nft logs end in journald. In order
to ensure we get a readable output, let's just grab the important
lines and create a dedicated file.
This will NOT work for iptables backend, since nftables is configured
with some specificities we can't set in iptables.
Since the log jump is at the very end of the chain, right before the
drop, anything logged is a packet that doesn't match any rule; we can
therefore understand them as "dropped packets", and would be able to
understand the actual reason, and add potential missing rules if
needed.
[1] https://review.opendev.org/c/openstack/tripleo-ansible/+/850222
Change-Id: I36d2fe7340c525c0438efc80bf73a103d90aa44d
nftables content will contain all of iptables - especially starting cs8,
where iptables is a compatibility wrapper for nftables (true name:
iptables-nft).
In addition, getting a dedicated file for nftables content makes things
easier to read, especially with the nftables output (think "json", more
or less).
Notes:
- iptables will still be called if the system can't find the
"nft" binary.
- this patch will be especially important once [1] get in, since
iptables will NOT see any of the nftables rules.
[1] https://review.opendev.org/c/openstack/tripleo-ansible/+/841414
Change-Id: Icba6b51ba5480091adcd2e010c9e34c049216c22
It may happen we need to get some more data out of OVN - let's gather
them.
This new output may help in case we have doubts about the actual OVN
config, especially its bridge and mappings.
Change-Id: I79ceee4d9b088d8b4dd5bd6935480fc05dd2ccf1
ansible-test seems to take the latest ansible-core by default, even if
we push upper/lower constraints in the requirements.txt or
test-requirements.txt.
Pushing the constraint directly in the tox.ini, for the "sanity"
testenv, seems to do the trick, as we're testing against an ansible
version we actually support, instead of the latest (2.13 today).
This patch also takes the opportunity to remove unwanted warnings
(ansible.cfg deprecated options), as well as ensuring the ignored
content for `ansible-test' doesn't raise any warning.
Closes-Bug: #1976251
Change-Id: I596aecd71ebe7d8a747ab8e6d05766f70dd78f1d
This can really help debugging issues, since we can see stats for
packets hitting the rules.
Take data out of the filter table only - but we may consider adding the
nat one if needed in the future.
Change-Id: If425a11de5a8ceaf8c3427dd14b0f135d3fa53ed
Setting desired version of black to 22.3.0 prevents issues with
the click dependency described in the #1966879.
Also modifies the docs/source/conf.py file to comply with
linting rules.
Related-Bug: #1966879
Signed-off-by: Jiri Podivin <jpodivin@redhat.com>
Change-Id: Icb5cc90d33ffdcb15e1870e424981d435683c77c
Tripleo-ci jobs are broken after latest release of setuptools 61.0,
details in Releated-Bug.
Trying the workaround as mentioned in [1]
[1] https://github.com/pypa/setuptools/issues/3197
Releated-Bug: #1966382
Change-Id: I02e08ab6225a1bdc897cdebda25365fc4865c716
The non-zuul-based reproducer has been unmaintained for years. Its
functionality is a strict subset of the zuul-based-reproducer
nowadays. Hence let's get rid of it.
Change-Id: If05b80b3a3cbf85d14e13d12af07d476a4b7b7b3
If IPA is installed, let's capture the config and dns configuration to
help troubleshoot ipa related things.
Change-Id: Ic45d32c46dbac2e81abde1a9116d34ffad14bf18
Related-Bug: #1957083
Add support for setting the artcl_find_maxdepth and artcl_find_max_size
variables via infrared's CLI when running the role as infrared plugin.
Change-Id: I3f5c26fe6ac5c69223c12d2574ce1659755c696c
Reason: when testing with FIPS enabled, collect logs for containers
is taking more to finish and ends timing out.
This patch updates async call timeout and async_status waiter tasks
to use the timeout in seconds defined by 'artcl_container_collect_timeout',
instead of a hardcoded value.
Change-Id: Ia936fa2894af5cf41c12a232aa0c1795d87400f7
Signed-off-by: Douglas Viroel <dviroel@redhat.com>
Those are needed to easier troubleshoot qemu issues
Change-Id: I3c605c471ffd8dac54d12a583f8c14146808e543
Signed-off-by: Bogdan Dobrelya <bdobreli@redhat.com>
The ternary is failing for some reason, so it is better to avoid the use
of ternary and split the task in two using the when: sova_config_file is
defined or not defined.
Change-Id: Ieb1ebe58608d2445c78942062cd9a82929ffdeb4
This adds the possibility to sova uses a local file instead of
downloading from github.
This is required because the verification searchs for a specific
failure, that recently was updated in the sova config file, and
the test started to fail. So, now, we use a static sova config
file to be used by the molecule, avoiding these kind of failures
happen in the future.
Change-Id: I2530188f1795a39ec6ec7e43c9ba118349c08541
Apparently, sealert doesn't allow this option to be used with other
options, and returns a wonderful:
"cannot run as noservice with other option"
Change-Id: I18d2a2efd21c2748a3e74907ba0654e51bf9fe82