[Webkit-unassigned] [Bug 186635] New: [GTK][buildbot] Raise timeouts for the step benchmark-test on the GTK perf bot

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=186635

            Bug ID: 186635
           Summary: [GTK][buildbot] Raise timeouts for the step
                    benchmark-test on the GTK perf bot
           Product: WebKit
           Version: WebKit Nightly Build
          Hardware: Unspecified
                OS: Unspecified
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: WebKit Gtk
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: clopez at igalia.com
                CC: bugs-noreply at webkitgtk.org

The step benchmark-test on the GTK perf bot runs all the benchmark plans one by one (one after the other).

Each one of this benchmarks plans have different timeouts.

The problem is that buildbot itself will kill the whole step when no output is produced after a 1200 seconds timeout, but some of the plans for run-benchmark have higher timeouts than 1200... that means that when things go wrong (or the machine/built-product is slower than expected) the buildbot timeout will happen before the plan one.

And when that happens, that means that the whole buildbot step will be stopped instead of having simply one test of the test set failing because its timeout has been reached.

This has been an already an issue on the past, and we already fixed this by growing the buildbot timeout for this step on the past


But it seems now some plans (like motionmark) have grown the plan timeout further, so we should adjust.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20180614/24fccc5f/attachment.html>