Closed MarSik closed 3 months ago
@MarSik: This pull request references Jira Issue OCPBUGS-34360, which is valid. The bug has been moved to the POST state.
Requesting review from QA contact: /cc @yliu127
The bug has been updated to refer to the pull request using the external bug tracker.
Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all
@MarSik: This pull request references Jira Issue OCPBUGS-34360, which is valid.
Requesting review from QA contact: /cc @yliu127
@ffromani Even the simple change makes quite a difference.
/lgtm
/assign davemulford
@sferich888 @davemulford Folks, who can provide the last missing approve here?
/lgtm
[APPROVALNOTIFIER] This PR is APPROVED
This pull-request has been approved by: ffromani, MarSik, sferich888
The full list of commands accepted by this bot can be found here.
The pull request process is described here
/retest-required
Remaining retests: 0 against base HEAD 408ab8661885cc7eb2140f71f9da9bef5f303805 and 2 for PR HEAD b952bd35c19f553fbe19aaf4c1116a1d8ad1a4b4 in total
/retest-required
@MarSik: all tests passed!
Full PR test history. Your PR dashboard.
@MarSik: Jira Issue OCPBUGS-34360: All pull requests linked via external trackers have merged:
Jira Issue OCPBUGS-34360 has been moved to the MODIFIED state.
/cherry-pick release-4.16
@MarSik: new pull request created: #426
[ART PR BUILD NOTIFIER]
This PR has been included in build ose-must-gather-container-v4.17.0-202406121512.p0.gb8be7da.assembly.stream.el9 for distgit ose-must-gather. All builds following this will include this PR.
The PPC node data collection was running in serial and that increased the time needed to collect the whole cluster. This change executes each per-node sequence in parallel and waits for all the sub tasks.
The only disadvantage is having a per-node process (potentially hundreds in huge clusters), however this case has always been lower in the code as well. The load on the controlling must-gather pod should be low as it only executes remote commands and waits for results.
This can be tested (at least while the PR is open) by executing:
Thanks to our scale testing colleagues we have a comparison between the stock and the parallelized runs using OCP 4.16.0-rc.3 on a cluster with 120 worker nodes.
stock:
with PR applied: