Closed
Bug 795649
Opened 12 years ago
Closed 12 years ago
Slave affinity should look for the last slave which successfully finished a job, not the last slave that finished it
Categories
(Release Engineering :: General, defect)
Release Engineering
General
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: philor, Assigned: catlee)
References
Details
In _getLastTimeOnBuilder in http://hg.mozilla.org/build/buildbotcustom/file/default/misc.py#l213, we look for the last time that a slave finished a particular job, no matter how it finished it, to give the next job to the slave that most recently finished it. In the case of a retriggered failed job because it looks like the slave might have been at fault, that's the last thing we want; in the case of a RETRY job because the slave is broken, we get strings of RETRY like https://tbpl.mozilla.org/?tree=Mozilla-Inbound&rev=b56f7cb51b1f&jobname=Android%20Tegra%20250%20mozilla-inbound%20opt%20test%20jsreftest-3 (currently on the 23rd try on the same slave which times out in verify.py).
We should instead be looking for the last time that slave did a job successfully. That would mean the search will be much slower for jobs which are permaorange, but frankly I don't much care about those jobs, unless they slow down scheduling for jobs worth doing, too.
Reporter | ||
Comment 1•12 years ago
|
||
Even simpler solution to the same problem (for tests, where it's mostly a problem): bug 714313
Depends on: 714313
Assignee | ||
Updated•12 years ago
|
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Updated•11 years ago
|
Product: mozilla.org → Release Engineering
Updated•7 years ago
|
Component: General Automation → General
You need to log in
before you can comment on or make changes to this bug.
Description
•