Run os_must_gather only in case of failure #457
Conversation
dasm
requested review from
cjeanner
and removed request for
Sandeepyadav93 and
psathyan
|
Build failed (check pipeline). Post https://review.rdoproject.org/zuul/buildset/548f96a749844e1fbb477dac6250fc2a ✔️ noop SUCCESS in 0s |
|
Just asking, there is no zuul fact set that tells us the result of the job? |
I'd rather avoid that, too zuul centric IMHO. |
dasm
force-pushed
the
log_gathering
branch
2 times, most recently
from
d40ca54
to
3d9e882
Compare
|
Build failed (check pipeline). Post https://review.rdoproject.org/zuul/buildset/7f905f3b8aea42908404287b302c0cdb ✔️ noop SUCCESS in 0s |
| rescue: | ||
| - name: Get CRC logs if os_must_gather failed | ||
| ansible.builtin.import_tasks: crc.yml | ||
| always: |
There was a problem hiding this comment.
IIUC @cjeanner you want to skip any logs collection on successful run.
I created a block around os_must_gather and crc.yml with edpm.yml . I hope it's going to allow us to spare a disk space and time execution as well.
There was a problem hiding this comment.
we probably don't need the always block for the edpm.yml inclusion.
But that's more cosmetics at this point. Should do the trick.
There was a problem hiding this comment.
If I wouldn't add always block, it couldn't be "disabled" when successful test run.
There was a problem hiding this comment.
hmm yep, by adding the same when condition :).
|
recheck |
|
Build failed (check pipeline). Post https://review.rdoproject.org/zuul/buildset/f872fc24777c4069bb83497e11c8399d ✔️ noop SUCCESS in 0s |
|
@cjeanner can you give me anouther round of review? |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: cjeanner The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
Run os_must_gather only in case of a failure which happened during "run" step in ZUUL. By doing so, we're reducing post-run job by about 450-600s, also preserving disk space.
|
/lgtm |
|
Build failed (check pipeline). Post https://review.rdoproject.org/zuul/buildset/85b52f2e5ed148e9bb3f85d9b67ac113 ✔️ noop SUCCESS in 0s |
|
recheck |
openshift-merge-robot
merged commit 54d80e3
into
openstack-k8s-operators:main
#457 enables running os_must_gather role in case of failure but also addes resuce task to run crc and edpm log collection. Since os_must_gather role runs must_gather tool to do the log collection and if must_gather fails, it also resuces operation to collect more logs on the node. The task associated with calling os_must_gather role is always going to get skipped if the job passes and we have no log collection for edpm and podified. Developers have no logs to verify/check few things in passed jobs. This pr fixes the same by running crc task file in case of passing job and os_must_gather role in case of failed job. We should always run edpm task files to collect edpm logs. Signed-off-by: Chandan Kumar <raukadah@gmail.com>
#457 enables running os_must_gather role in case of failure but also addes resuce task to run crc and edpm log collection. Since os_must_gather role runs must_gather tool to do the log collection and if must_gather fails, it also resuces operation to collect more logs on the node. The task associated with calling os_must_gather role is always going to get skipped if the job passes and we have no log collection for edpm and podified. Developers have no logs to verify/check few things in passed jobs. This pr fixes the same by running crc task file in case of passing job and os_must_gather role in case of failed job. We should always run edpm task files to collect edpm logs. Signed-off-by: Chandan Kumar <raukadah@gmail.com>
#457 enables running os_must_gather role in case of failure but also addes resuce task to run crc and edpm log collection. Since os_must_gather role runs must_gather tool to do the log collection and if must_gather fails, it also resuces operation to collect more logs on the node. The task associated with calling os_must_gather role is always going to get skipped if the job passes and we have no log collection for edpm and podified. Developers have no logs to verify/check few things in passed jobs. This pr fixes the same by running crc task file in case of passing job and os_must_gather role in case of failed job. We should always run edpm task files to collect edpm logs. Signed-off-by: Chandan Kumar <raukadah@gmail.com>
Run os_must_gather only in case of a failure which happened during "run" step in ZUUL.
By doing so, we're reducing post-run job by about 450-600s, also preserving disk space.
As a pull request owner and reviewers, I checked that: