You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@pulsar.apache.org by GitBox <gi...@apache.org> on 2021/02/04 11:15:42 UTC
[GitHub] [pulsar] lhotari opened a new issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
lhotari opened a new issue #9479:
URL: https://github.com/apache/pulsar/issues/9479
Cancelling workflows with airflow-cancel-workflow-runs within each workflow doesn't seem to work. That change was made in #9357 .
However there's something that is cancelling workflows. Cancelling happened for example for PR #9443 . All of it's checks are now in cancelled state. I checked also a few other cases. It looks like someone is manually maintaining the build queue and cancelling jobs.
Another example I looked at was #9476 . The remaining jobs from the previous commit had been already cancelled even though the new jobs hadn't yet started. Currently the automated cancellation will only happen after the next workflow run has started.
This issue report is about errors seen for in the logs of this workflow run:
https://github.com/apache/pulsar/runs/1804639066#step:3:119
```
2021-02-01T09:09:48.6152871Z Cancelling run: 527073250.
2021-02-01T09:09:48.6153130Z
2021-02-01T09:09:48.7047449Z ##[warning]
Could not cancel run 527073250: [403] Resource not accessible by integration
2021-02-01T09:09:48.7062237Z
2021-02-01T09:09:48.7062636Z Cancelling run: 125656974.
2021-02-01T09:09:48.7062888Z
2021-02-01T09:09:48.8020647Z ##[warning]
Could not cancel run 125656974: [403] Resource not accessible by integration
2021-02-01T09:09:48.8022156Z
2021-02-01T09:09:48.8022528Z Cancelling run: 124625158.
2021-02-01T09:09:48.8022809Z
2021-02-01T09:09:48.8864785Z ##[warning]
Could not cancel run 124625158: [403] Resource not accessible by integration
```
<details>
<summary>Full logs for the cancellation step</summary>
<code><pre>
2021-02-01T09:09:47.0409051Z ##[group]Run apache/airflow-cancel-workflow-runs@953e057dc81d3458935a18d1184c386b0f6b5738
2021-02-01T09:09:47.0410182Z with:
2021-02-01T09:09:47.0411786Z token: ***
2021-02-01T09:09:47.0412340Z env:
2021-02-01T09:09:47.0413416Z MAVEN_OPTS: -Dhttp.keepAlive=false -Dmaven.wagon.http.pool=false -Dmaven.wagon.http.retryHandler.count=3
2021-02-01T09:09:47.0414822Z ##[endgroup]
2021-02-01T09:09:47.4123083Z [@octokit/rest] `const Octokit = require("@octokit/rest")` is deprecated. Use `const { Octokit } = require("@octokit/rest")` instead
2021-02-01T09:09:47.4379824Z [@octokit/rest] `const Octokit = require("@octokit/rest")` is deprecated. Use `const { Octokit } = require("@octokit/rest")` instead
2021-02-01T09:09:47.4451574Z
2021-02-01T09:09:47.4452863Z Getting workflow id for source run id: 526558278, owner: apache, repo: pulsar, skipEventTypes:
2021-02-01T09:09:47.4453418Z
2021-02-01T09:09:47.5954282Z The source run 526558278 is in https://api.github.com/repos/apache/pulsar/actions/workflows/955 workflow
2021-02-01T09:09:47.5955439Z
2021-02-01T09:09:47.5956145Z Finding runs for my own workflow 955
2021-02-01T09:09:47.5956718Z
2021-02-01T09:09:47.5957960Z Repository: apache/pulsar, Owner: apache, Repo: pulsar, Event name: pull_request, CancelMode: duplicates, sourceWorkflowId: 955, sourceRunId: 526558278, selfRunId: 526558278, jobNames:
2021-02-01T09:09:47.7072108Z Source workflow: Head repo: aloyszhang/pulsar, Head branch: fix Event: pull_request, Head sha: dd34a5b547f194d3d03fdc6d798ed7dba27d5124, url: https://api.github.com/repos/apache/pulsar/actions/runs/526558278
2021-02-01T09:09:47.7073359Z
2021-02-01T09:09:47.7074041Z Finding PR request id for: owner: apache, Repo:pulsar, Head:aloyszhang:fix.
2021-02-01T09:09:47.7074592Z
2021-02-01T09:09:47.8817175Z
2021-02-01T09:09:47.8818545Z Comparing: 9334 sha: dd34a5b547f194d3d03fdc6d798ed7dba27d5124 with expected: dd34a5b547f194d3d03fdc6d798ed7dba27d5124.
2021-02-01T09:09:47.8825915Z
2021-02-01T09:09:47.8826135Z
2021-02-01T09:09:47.8826514Z Found PR: 9334
2021-02-01T09:09:47.8826742Z
2021-02-01T09:09:47.8827256Z Setting output: sourceHeadRepo: aloyszhang/pulsar
2021-02-01T09:09:47.8838946Z Setting output: sourceHeadBranch: fix
2021-02-01T09:09:47.8841018Z Setting output: sourceHeadSha: dd34a5b547f194d3d03fdc6d798ed7dba27d5124
2021-02-01T09:09:47.8842236Z Setting output: sourceEvent: pull_request
2021-02-01T09:09:47.8843303Z Setting output: pullRequestNumber: 9334
2021-02-01T09:09:47.8844500Z Setting output: mergeCommitSha: 7da506c8190ed5aebbd8f78cacdf357865a9d3f6
2021-02-01T09:09:47.8846276Z Setting output: targetCommitSha: 7da506c8190ed5aebbd8f78cacdf357865a9d3f6
2021-02-01T09:09:47.8847247Z
2021-02-01T09:09:47.8847619Z ###################################################################################
2021-02-01T09:09:47.8847889Z
2021-02-01T09:09:47.8849102Z All parameters: owner: apache, repo: pulsar, run id: 526558278, Source workflow id: 955, head repo aloyszhang/pulsar, headBranch: fix, sourceEventName: pull_request, cancelMode: duplicates, jobNames:
2021-02-01T09:09:47.8849991Z
2021-02-01T09:09:47.8850879Z ###################################################################################
2021-02-01T09:09:47.8851138Z
2021-02-01T09:09:47.8851850Z # Cancel duplicate runs for workflow 955 for same triggering branch as own run Id.
2021-02-01T09:09:47.8852321Z
2021-02-01T09:09:47.8853069Z ###################################################################################
2021-02-01T09:09:47.8853414Z
2021-02-01T09:09:47.8853620Z
2021-02-01T09:09:47.8854350Z Finding duplicate runs: Owner: apache, Repo: pulsar, Status: queued Workflow ID:955, Head Branch: fix,Event name: pull_request
2021-02-01T09:09:47.8854985Z
2021-02-01T09:09:48.4512396Z
2021-02-01T09:09:48.4513555Z Finding duplicate runs: Owner: apache, Repo: pulsar, Status: in_progress Workflow ID:955, Head Branch: fix,Event name: pull_request
2021-02-01T09:09:48.4514183Z
2021-02-01T09:09:48.6103631Z
2021-02-01T09:09:48.6104310Z Found runs: 7212,7211,2654,2610,7192
2021-02-01T09:09:48.6111163Z
2021-02-01T09:09:48.6111472Z
2021-02-01T09:09:48.6113228Z Checking run number: 7212 RunId: 527117669 Url: https://api.github.com/repos/apache/pulsar/actions/runs/527117669 Status queued Created at 2021-02-01T08:34:21Z
2021-02-01T09:09:48.6113940Z
2021-02-01T09:09:48.6114212Z
2021-02-01T09:09:48.6115042Z The run 527117669 is from a different repo: eolivelli/pulsar (expected aloyszhang/pulsar). Not adding as candidate to cancel.
2021-02-01T09:09:48.6115935Z
2021-02-01T09:09:48.6116140Z
2021-02-01T09:09:48.6116711Z Cancel Future Duplicates: Returning run id that might be duplicate or my own run: 527117669.
2021-02-01T09:09:48.6117207Z
2021-02-01T09:09:48.6117410Z
2021-02-01T09:09:48.6118359Z Adding the run: 527117669 to candidates with :955:aloyszhang/pulsar:fix:pull_request key.
2021-02-01T09:09:48.6118845Z
2021-02-01T09:09:48.6119177Z
2021-02-01T09:09:48.6120496Z Checking run number: 7211 RunId: 527073250 Url: https://api.github.com/repos/apache/pulsar/actions/runs/527073250 Status queued Created at 2021-02-01T08:14:22Z
2021-02-01T09:09:48.6121122Z
2021-02-01T09:09:48.6121303Z
2021-02-01T09:09:48.6121954Z The run 527073250 is from a different repo: BewareMyPower/pulsar (expected aloyszhang/pulsar). Not adding as candidate to cancel.
2021-02-01T09:09:48.6122535Z
2021-02-01T09:09:48.6122713Z
2021-02-01T09:09:48.6123675Z Cancel Future Duplicates: Returning run id that might be duplicate or my own run: 527073250.
2021-02-01T09:09:48.6124372Z
2021-02-01T09:09:48.6124696Z
2021-02-01T09:09:48.6125452Z Adding the run: 527073250 to candidates with :955:aloyszhang/pulsar:fix:pull_request key.
2021-02-01T09:09:48.6125937Z
2021-02-01T09:09:48.6126130Z
2021-02-01T09:09:48.6127304Z Checking run number: 2654 RunId: 125656974 Url: https://api.github.com/repos/apache/pulsar/actions/runs/125656974 Status queued Created at 2020-06-05T06:35:30Z
2021-02-01T09:09:48.6128261Z
2021-02-01T09:09:48.6128451Z
2021-02-01T09:09:48.6129166Z The run 125656974 is from a different repo: magrain/pulsar (expected aloyszhang/pulsar). Not adding as candidate to cancel.
2021-02-01T09:09:48.6129776Z
2021-02-01T09:09:48.6130077Z
2021-02-01T09:09:48.6130657Z Cancel Future Duplicates: Returning run id that might be duplicate or my own run: 125656974.
2021-02-01T09:09:48.6131333Z
2021-02-01T09:09:48.6131542Z
2021-02-01T09:09:48.6132105Z Adding the run: 125656974 to candidates with :955:aloyszhang/pulsar:fix:pull_request key.
2021-02-01T09:09:48.6132593Z
2021-02-01T09:09:48.6132783Z
2021-02-01T09:09:48.6133956Z Checking run number: 2610 RunId: 124625158 Url: https://api.github.com/repos/apache/pulsar/actions/runs/124625158 Status queued Created at 2020-06-04T09:04:51Z
2021-02-01T09:09:48.6134672Z
2021-02-01T09:09:48.6134864Z
2021-02-01T09:09:48.6135571Z The run 124625158 is from a different repo: magrain/pulsar (expected aloyszhang/pulsar). Not adding as candidate to cancel.
2021-02-01T09:09:48.6136185Z
2021-02-01T09:09:48.6136374Z
2021-02-01T09:09:48.6136959Z Cancel Future Duplicates: Returning run id that might be duplicate or my own run: 124625158.
2021-02-01T09:09:48.6137602Z
2021-02-01T09:09:48.6137797Z
2021-02-01T09:09:48.6138370Z Adding the run: 124625158 to candidates with :955:aloyszhang/pulsar:fix:pull_request key.
2021-02-01T09:09:48.6138853Z
2021-02-01T09:09:48.6139042Z
2021-02-01T09:09:48.6140199Z Checking run number: 7192 RunId: 526558278 Url: https://api.github.com/repos/apache/pulsar/actions/runs/526558278 Status in_progress Created at 2021-02-01T03:35:40Z
2021-02-01T09:09:48.6140944Z
2021-02-01T09:09:48.6141132Z
2021-02-01T09:09:48.6142212Z I have self-preservation built in. I refuse to cancel myself :)
2021-02-01T09:09:48.6142811Z
2021-02-01T09:09:48.6148088Z
2021-02-01T09:09:48.6149541Z Skipping the first run (527117669) of all the matching duplicates for ':955:aloyszhang/pulsar:fix:pull_request'. This one we are going to leave in peace!
2021-02-01T09:09:48.6150278Z
2021-02-01T09:09:48.6150459Z
2021-02-01T09:09:48.6151242Z ###### Cancelling runs for :955:aloyszhang/pulsar:fix:pull_request starting from the most recent ##########
2021-02-01T09:09:48.6151746Z
2021-02-01T09:09:48.6152095Z Number of runs to cancel: 3
2021-02-01T09:09:48.6152366Z
2021-02-01T09:09:48.6152542Z
2021-02-01T09:09:48.6152871Z Cancelling run: 527073250.
2021-02-01T09:09:48.6153130Z
2021-02-01T09:09:48.7047449Z ##[warning]
Could not cancel run 527073250: [403] Resource not accessible by integration
2021-02-01T09:09:48.7062237Z
2021-02-01T09:09:48.7062636Z Cancelling run: 125656974.
2021-02-01T09:09:48.7062888Z
2021-02-01T09:09:48.8020647Z ##[warning]
Could not cancel run 125656974: [403] Resource not accessible by integration
2021-02-01T09:09:48.8022156Z
2021-02-01T09:09:48.8022528Z Cancelling run: 124625158.
2021-02-01T09:09:48.8022809Z
2021-02-01T09:09:48.8864785Z ##[warning]
Could not cancel run 124625158: [403] Resource not accessible by integration
2021-02-01T09:09:48.8866064Z
2021-02-01T09:09:48.8866581Z ###### Finished cancelling runs for :955:aloyszhang/pulsar:fix:pull_request ##########
2021-02-01T09:09:48.8867216Z
2021-02-01T09:09:48.8867667Z Setting output: cancelledRuns: [527073250,125656974,124625158]
2021-02-01T09:09:48.8868918Z
2021-02-01T09:09:48.8869475Z ############### Cancel complete ##################
</pre></code>
</details>
It seems that airflow-cancel-workflow-runs is not supposed to be used in the way that we use it.
There are actions that support per-workflow cancelling, for example https://github.com/marketplace/actions/cancel-workflow-action . That was previously used, until it was possibly unintentionally removed by #9069 .
It would have been possible to continue to use cancel-workflow-action, as described in https://github.com/apache/pulsar/issues/9154#issuecomment-756984731
> An alternative option is to raise INFRA ticket to allow styfle/cancel-workflow-action (and still use SHA-based action ids like styfle/cancel-workflow-action@123123123...123123)
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lhotari edited a comment on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
lhotari edited a comment on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773242299
@potiuk I guess airflow-cancel-workflow-runs would be perfect if we didn't have so many workflows. I wonder if it would be possible to have it get triggered only by a single workflow request, but still cancel all runs in all workflows for that particular branch (in pull requests, the pr branch)? That could solve the issue that the build queue gets filled by the cancellation jobs since there would be only 1 cancellation job instead of 25.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] potiuk commented on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
potiuk commented on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773535003
You are inded usign the action wrongly.
The problem is that the workflow that runs from fork does not have permission to write in your repo:
This is what `Could not cancel run 527073250: [403] Resource not accessible by integration` message means (thought it is not obvious).
The actual "cancel" action shoudl run in a separate worklfow (of `workflow_run` type): https://docs.github.com/en/actions/reference/events-that-trigger-workflows#workflow_run
If you start many workflows at the same time, it will be enough to start single "workflow_run" with multiple cancel-workflow steps (one per-workflow).
I think also you misuderstand how allDuplicates action works. It does not kill "one branch" duplicates" - it is far more aggressive, It kills ALL duplicates in ALL branches for a given workflow.
I can - tomorrow- draft a PR fixing your setup if you would like my help with that :)
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] potiuk commented on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
potiuk commented on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773535003
You are inded usign the action wrongly.
The problem is that the workflow that runs from fork does not have permission to write in your repo:
This is what `Could not cancel run 527073250: [403] Resource not accessible by integration` message means (thought it is not obvious).
The actual "cancel" action shoudl run in a separate worklfow (of `workflow_run` type): https://docs.github.com/en/actions/reference/events-that-trigger-workflows#workflow_run
If you start many workflows at the same time, it will be enough to start single "workflow_run" with multiple cancel-workflow steps (one per-workflow).
I think also you misuderstand how allDuplicates action works. It does not kill "one branch" duplicates" - it is far more aggressive, It kills ALL duplicates in ALL branches for a given workflow.
I can - tomorrow- draft a PR fixing your setup if you would like my help with that :)
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] potiuk edited a comment on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
potiuk edited a comment on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773535003
You are inded usign the action wrongly.
The problem is that the workflow that runs from fork does not have permission to write in your repo:
This is what `Could not cancel run 527073250: [403] Resource not accessible by integration` message means (thought it is not obvious).
The actual "cancel" action shoudl run in a separate worklfow (of `workflow_run` type): https://docs.github.com/en/actions/reference/events-that-trigger-workflows#workflow_run - this way it will have WRITE permission necessary to cancel workflows.
If you start many workflows at the same time, it will be enough to start single "workflow_run" with multiple cancel-workflow steps (one per-workflow).
I think also you misuderstand how allDuplicates action works. It does not kill "one branch" duplicates" - it is far more aggressive, It kills ALL duplicates in ALL branches for a given workflow.
I can - tomorrow- draft a PR fixing your setup if you would like my help with that :)
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lhotari commented on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
lhotari commented on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773242299
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] potiuk edited a comment on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
potiuk edited a comment on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773535003
You are inded usign the action wrongly.
The problem is that the workflow that runs from fork does not have permission to write in your repo:
This is what `Could not cancel run 527073250: [403] Resource not accessible by integration` message means (thought it is not obvious).
The actual "cancel" action shoudl run in a separate worklfow (of `workflow_run` type): https://docs.github.com/en/actions/reference/events-that-trigger-workflows#workflow_run - this way it will have WRITE permission necessary to cancel workflows.
If you start many workflows at the same time, it will be enough to start single "workflow_run" with multiple cancel-workflow steps (one per-workflow).
I think also you misuderstand how allDuplicates action works. It does not kill "one branch" duplicates" - it is far more aggressive, It kills ALL duplicates in ALL branches for a given workflow.
I can - tomorrow- draft a PR fixing your setup if you would like my help with that :)
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lhotari commented on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
lhotari commented on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773242299
@potiuk I guess airflow-cancel-workflow-runs would be perfect if we didn't have so many workflows. I wonder if it would be possible to have it get triggered only by a single workflow request, but still cancel all runs in all workflows for that particular branch?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lhotari commented on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
lhotari commented on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773819768
> I think also you misuderstand how allDuplicates action works. It does not kill "one branch" duplicates" - it is far more aggressive, It kills ALL duplicates in ALL branches for a given workflow.
@potiuk Is there a mode where one could limit "killing of the duplicates" to the same branch as the new triggered workflow run?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] potiuk commented on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
potiuk commented on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773985961
> @potiuk Is there a mode where one could limit "killing of the duplicates" to the same branch as the new triggered workflow run?
Yes. The `duplicates` mode works this way (see docs below). It was the original mode it worked on by default but it is not really effective in the "hight strain' situation as it turned out, and that's why I introduced `allDuplicates`.
Why this is not a good idea to use `duplicates` while we are having the "high queue strain" problem?
Because of the way how workflows are queued. The `allDuplicates` mode have been introduced precisely to tackle the "high queue strain" problem. Problem with using the action is that it has to actually start running to cancel anything. And if you have a "strain" on the GitHub Action Queue, your "cancel" workflow might simply be queued for hours because other jobs will occupy the queue. So the past "duplicates" from the same branch will not be cancelled effectively because the job to cancel them will only run hours later after those duplicates have finished.
If we really want to be "queue conscious", I honestly cannot imagine a situation where you would like to wait for previous build from the same branch to complete when you already have a new one queued. Yeah you will see the result later, but there is probably a reason why you pushed new changes to that branch and what you are really interested in eventually is to see the results of the build for the "newest" commit in that branch, not the previous ones. And yeah - it means that you might see some partial results of it later, but this is what I consider as being "nice" to others. When you push new version of the same PR, cancelling any (even already running) version of that PR is simply a very nice gesture for the others, because you will anyhow have to look at the result of the newer commits.
That's why the `allDuplicates` is much better approach IMHO. It is more aggressive, yes, but also it is much more efficient. Whenever ANY cancel jobs will run, it will cancel ALL duplicates from ALL branches there - therefore it makes all people who push more than one change successively pushing changes to their PRs to ONLY use the job time for the last version of the commit they pushed (in the "strain" situation). I think this is really the best way to proceed.
Here, relevant part from the original documentation from https://github.com/potiuk/cancel-workflow-runs
The job cancel modes work as follows:
Cancel Mode | No sourceRunId specified | The sourceRunId set to ${{ github.event.workflow_run.id }}
-- | -- | --
duplicates | Cancels duplicate runs from the same repo/branch as current run. | Cancels duplicate runs for the same repo/branch as the source run.
allDuplicates | Cancels duplicate runs from all running workflows. | Cancels duplicate runs from all running workflows.
self | Cancels self run. | Cancel the sourceRunId run.
failedJobs | Cancels all runs of own workflow that have matching jobs that failed. | Cancels all runs of the sourceRunId workflow that have matching jobs that failed.
namedJobs | Cancels all runs of own workflow that have matching jobs. | Cancels all runs of the sourceRunId workflow that have matching jobs.
allDuplicatedNamedJobs | Cancels all duplicate runs of own workflow that share matching jobs pattern. | Cancels all runs of the sourceRunId workflow that share matching job pattern.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lhotari edited a comment on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
lhotari edited a comment on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773242299
@potiuk I guess airflow-cancel-workflow-runs would be perfect if we didn't have so many workflows. I wonder if it would be possible to have it get triggered only by a single workflow request, but still cancel all runs in all workflows for that particular branch? That could solve the issue that the build queue gets filled by the cancellation jobs since there would be only 1 cancellation job instead of 25.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lhotari commented on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
lhotari commented on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-775176225
Fixed by @potiuk 's PR #9503 which has been merged.
@potiuk Thank you for the great help!
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] potiuk commented on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
potiuk commented on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-774470368
Sorry it's been a little later PR is here: https://github.com/apache/pulsar/pull/9503
Since it need master merge first, I run it on my fork to test it here:
https://github.com/potiuk/pulsar/runs/1844795428
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] potiuk edited a comment on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
potiuk edited a comment on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773985961
> @potiuk Is there a mode where one could limit "killing of the duplicates" to the same branch as the new triggered workflow run?
Yes. The `duplicates` mode works this way (see docs below). It was the original mode it worked on by default but it is not really effective in the "high strain" situation as it turned out, and that's why I introduced `allDuplicates`.
Why this is not a good idea to use `duplicates` while we are having the "high queue strain" problem?
Because of the way how workflows are queued. The `allDuplicates` mode have been introduced precisely to tackle the "high queue strain" problem. Problem with using the action is that it has to actually start running to cancel anything. And if you have a "strain" on the GitHub Action Queue, your "cancel" workflow might simply be queued for hours because other jobs will occupy the queue. So the past "duplicates" from the same branch will not be cancelled effectively because the job to cancel them will only run hours later after those duplicates have finished.
If we really want to be "queue conscious", I honestly cannot imagine a situation where you would like to wait for previous build from the same branch to complete when you already have a new one queued. Yeah you will see the result later, but there is probably a reason why you pushed new changes to that branch and what you are really interested in eventually is to see the results of the build for the "newest" commit in that branch, not the previous ones. And yeah - it means that you might see some partial results of it later, but this is what I consider as being "nice" to others. When you push new version of the same PR, cancelling any (even already running) version of that PR is simply a very nice gesture for the others, because you will anyhow have to look at the result of the newer commits.
That's why the `allDuplicates` is much better approach IMHO. It is more aggressive, yes, but also it is much more efficient. Whenever ANY cancel jobs will run, it will cancel ALL duplicates from ALL branches there - therefore it makes all people who push more than one change successively pushing changes to their PRs to ONLY use the job time for the last version of the commit they pushed (in the "strain" situation). I think this is really the best way to proceed.
Here, relevant part from the original documentation from https://github.com/potiuk/cancel-workflow-runs
The job cancel modes work as follows:
Cancel Mode | No sourceRunId specified | The sourceRunId set to ${{ github.event.workflow_run.id }}
-- | -- | --
duplicates | Cancels duplicate runs from the same repo/branch as current run. | Cancels duplicate runs for the same repo/branch as the source run.
allDuplicates | Cancels duplicate runs from all running workflows. | Cancels duplicate runs from all running workflows.
self | Cancels self run. | Cancel the sourceRunId run.
failedJobs | Cancels all runs of own workflow that have matching jobs that failed. | Cancels all runs of the sourceRunId workflow that have matching jobs that failed.
namedJobs | Cancels all runs of own workflow that have matching jobs. | Cancels all runs of the sourceRunId workflow that have matching jobs.
allDuplicatedNamedJobs | Cancels all duplicate runs of own workflow that share matching jobs pattern. | Cancels all runs of the sourceRunId workflow that share matching job pattern.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lhotari commented on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
lhotari commented on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773578932
Thank you @potiuk , that would be helpful.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lhotari closed issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
lhotari closed issue #9479:
URL: https://github.com/apache/pulsar/issues/9479
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [pulsar] lhotari edited a comment on issue #9479: [Build] Cancelling workflows with airflow-cancel-workflow-runs action within each workflow doesn't seem to work
Posted by GitBox <gi...@apache.org>.
lhotari edited a comment on issue #9479:
URL: https://github.com/apache/pulsar/issues/9479#issuecomment-773242299
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org