The error is on branch-heads/2892 (log link https://build.chromium.org/p/chromium.infra.cron/builders/gnumbd/builds/74237/steps/gnumbd/logs/stdio).
This is likely due to failed branching, as this is the latest branch.
However, the failure itself is a because pending-tags/X contains a commit that isn't in pending/X, so either there was a incorrect fix attempt or there was some kind of race in pushing these refs.


[E2016-10-17T09:23:31.207297+00:00 30339 140631467784000 infra.services.gnumbd.gnumbd:458] Ref(Repo('https://chromium.googlesource.com/chromium/src'), 'refs/pending-tags/branch-heads/2892') and Ref(Repo('https://chromium.googlesource.com/chromium/src'), 'refs/pending/branch-heads/2892') match, but Ref(Repo('https://chromium.googlesource.com/chromium/src'), 'refs/branch-heads/2892')'s content doesn't match (diff {'footer_lines': "(('Review-Url', 'https://codereview.chromium.org/2420933002'),) != (('Review-Url', 'https://codereview.chromium.org/2417773003'),)", 'tree': 'df29da95b4ab8dbb2523488e8c061dc6b43c74e3 != afa688105acb233e9d42976ca2c6539f7e3a9364', 'message_lines': '(\'predictors: Add trace events to the resource_prefetch_predictor.\', \'\', \'BUG=631966\') != ("Fix content upside down when call OffscreenCanvas\'s commit", \'\', \'The reason that the content is upside down is that we have a flag called\', \'yflipped and it is always false. In the case of GPU compositing, this flag\', \'should be true.\', \'\', "This CL also updates pixel tests for WebGL\'s commit, such that instead", \'of drawing a uniform color square, we draw a white triangle on a green\', \'background.\', \'\', \'BUG= 655333 ,  655311 \', \'CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.android:android_optional_gpu_tests_rel\')', 'author': 'lizeb <lizeb@chromium.org> 1476635299 -0700 != xidachen <xidachen@chromium.org> 1476558999 -0700'})!

While builder is red, it's actually processing all other branches just fine. So, tree needs not be closed on it, nor is this P0.

This looks like a bad branch to me. Going to make refs/branch-heads/2892 match the pending one (which is refs/heads/master@{#425592})

This should be fixed on the next gnumbd build.

Yep, fixed.