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

Issue 687224 link

Starred by 1 user

Issue metadata

Status: Archived
Owner:
Closed: Feb 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

unexpectedly low build numbers in viceroy graph

Project Member Reported by semenzato@chromium.org, Jan 31 2017

Issue description

If you look at the small graph in the top right showing the duration of recent builds, for instance here:

https://viceroy.corp.google.com/chromeos/build_details?build_config=caroline-release&build_number=328&experimental=False&duration=1d&refresh=-1

there are strangely low build numbers in the mix.  Most builds are in the 300s, but a few are below 100, for instance this one:

caroline-release
Build number	< 56 > >>
Build hostname	cros-beefy179-c2.c.chromeos-bot.internal
Start time	Jan. 15, 2017, 9:03 p.m.
Finish time	Jan. 16, 2017, 2:31 a.m.
Duration	19707 seconds
Status	pass

Where are these strange build numbers coming from?

 
Cc: akes...@chromium.org pho...@chromium.org
These are release builds on another branch/waterfall.  If you look at the build with 56 you'll see the builder name and waterfall indicate that.

When I originally built things I didn't have a good enough mental model and built things assuming build_config + build number uniquely identify a build.  I think if we change it to waterfall/build_config/build_number, while still not guaranteed to be unique (buildbot_generation could technically change and roll things over), it would be sufficient to address these issues.
Cc: jinjingl@chromium.org
+jinjingl who might need to deal with this for GoldenEye as well.
Thank you for the clarification, David.

If it's the same kind of build, just a different branch, maybe it's fine to leave it in the mix?  I don't suppose that the branch is likely to diverge much from ToT in terms of build time.

I am almost tempted to call this a WontFix, but I'll leave it to you.
Fix posted at cl/146135697.

If a waterfall or milestone version are specified, those filters will be applied.  Waterfall will default to 'chromeos' unless milestone version is specified.  

The most recent builds will apply the same logic, except instead of default to chromeos, it will default to the waterfall of the build being displayed.


Re: c#3 I've cleaned it up to try and remove the confusion.  I think there's still some edge cases that might not be handled 100% correctly, but it's also not clear what the absolute correct behaviour is, so I think I'm okay with leaving those edge cases unhandled for now.
Status: Fixed (was: Untriaged)

Comment 6 by dchan@google.com, Apr 17 2017

Labels: VerifyIn-59

Comment 7 by dchan@google.com, May 30 2017

Labels: VerifyIn-60

Comment 8 by dchan@chromium.org, Aug 1 2017

Labels: VerifyIn-61

Comment 9 by dchan@chromium.org, Oct 14 2017

Status: Archived (was: Fixed)

Sign in to add a comment