New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 682285 link

Starred by 1 user

Issue metadata

Status: Archived
Owner:
Last visit > 30 days ago
Closed: Jun 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

VersionNotSetError: builder must call SetVersionInfo first in lumpy-incremental-paladin

Project Member Reported by semenzato@chromium.org, Jan 18 2017

Issue description

https://uberchromegw.corp.google.com/i/chromeos/builders/lumpy-incremental-paladin/builds/10967/steps/BuildPackages%20%28pre-Patch%29/logs/stdio

The paladin has failed the last 5 runs in the same stage.  I suspect it's in some bad state, but just in case here are the CLs common to all 5 runs.

      5 portage-stable | vapier | 428970
      5 portage-stable | vapier | 428950
      5 platform2 | vapier | 428951
      5 firmware | hungte | 428513
      5 autotest | briannorris | 428551




 
Cc: vapier@chromium.org
Mike, you may know what goes on here.
Cc: nxia@chromium.org chingcodes@chromium.org
Adding more people who worked on related code recently.
semenzato@luigi:~/.../chromite/cbuildbot/stages$ git show 9a58098e8

This shows that a change from 12/17 is using self.archive_path, which was not used before.  That member is a "property", that is it causes the invocation of user-defined methods.  The failure occurs in one of that methods.



Comment 4 by dshi@chromium.org, Jan 18 2017

Owner: chingcodes@chromium.org
CL reverted
https://chromium-review.googlesource.com/#/c/430110/

Comment 5 by dshi@chromium.org, Jan 18 2017

Labels: -Pri-0 Pri-2
Could you elaborate what the issue is with using the archive_path?

Looking into the builds more..
Hmm. Yep, that makes sense.

chingcodes@ on the incremental builder, we run BuildPackages twice : once with the tree prior to applying test patches, and once after. During that first phase, cbuildbot doesn't yet know the version name of the build that we are testing (that gets defined later on in CommitQueueSyncStage) and so apparently self.archive_path is therefore undefined.

I think you should modify your logic to catch the VersionNotSetError there if it is raised, and if so simply not do the build event logging. We can leave a TODO behind to untangle the problem better eventually.
Ok, added quick catch of VersionNotsetError.

https://chromium-review.googlesource.com/#/c/430092/
Status: Fixed (was: Untriaged)

Comment 10 by dchan@chromium.org, Jan 22 2018

Status: Archived (was: Fixed)

Sign in to add a comment