691761 - Terminate jobs early for certain error conditions

Status: RESOLVED WORKSFORME

(Release Engineering :: General, defect, P5)

Description

[Joey Armstrong [:joey]]

Certain system/host failures should be allowed to terminate submissions early.
Two problems were encountered recently and one manifested downstream in l10n/packaging which resulted in a patch being backed out for unrelated problems.

Disk/filesystem issues should be one condition for automatic failure.  Transient errors are not fun to diagnose:
    https://bugzilla.mozilla.org/show_bug.cgi?id=690337

Checkout failures, disk full, incomplete sandboxes are another:
  abort: There is no Mercurial repository here (.hg not found)!
    https://bugzilla.mozilla.org/show_bug.cgi?id=680246
    https://bugzilla.mozilla.org/show_bug.cgi?id=554752


command: START
command: hg pull -r a2762135831ef473ddc590cceaa63c1fffeebab9 http://hg.mozilla.org/mozilla-central
command: cwd: e:\builds\moz2_slave\m-cen-w32\build
command: output:
abort: There is no Mercurial repository here (.hg not found)!
command: END (0.16s elapsed)

Comment 1

[Joey Armstrong [:joey]]

Wrong bug, 554752 should be 691753

https://bugzilla.mozilla.org/show_bug.cgi?id=691753

Comment 2

[Chris Cooper [:coop] (he/him)]

We added these changes elsewhere, and also started using hgtool.py (and upgrade hg) to avoid some of the cloning issues.