If you think a bug might affect users in the 57 release, please set the correct tracking and status flags for Release Management.
Bug 876526 (tst-linux64-ec2-390)

tst-linux64-ec2-390 problem tracking

RESOLVED FIXED

Status

Release Engineering
Buildduty
P3
normal
RESOLVED FIXED
4 years ago
4 years ago

People

(Reporter: philor, Unassigned)

Tracking

Firefox Tracking Flags

(Not tracked)

Details

(Whiteboard: [buildduty][buildslaves][capacity])

(Reporter)

Description

4 years ago
The boy ain't right: not only does it disconnect far too frequently, but https://tbpl.mozilla.org/php/getParsedLog.php?id=23452862&tree=Mozilla-Inbound was a run where it was running so slowly that it timed out a mochitest that never times out, and then took the screenshot so slowly and output the data: URI so slowly that the next test to run was interleaving its output (which is a horrible thing to find in the middle of a broken screenshot URI, since it outputs "@@@@@@@@@hi there" several times).

Nick says he disabled it.
Terminated the VM. 

Rail, do we recreate these at all ? What's involved in that ? I could probably figure out how to modify https://wiki.mozilla.org/ReleaseEngineering/How_To/Create_AWS_staging_slaves but I'm not sure how many puppet masters would need to forget the existing key.
Flags: needinfo?(rail)
Put this in the 'decomm' environment in slavealloc.
(In reply to nthomas@mozilla.com from comment #1)
> Terminated the VM. 
> 
> Rail, do we recreate these at all ? What's involved in that ?

Not right away, but once in 2-4 weeks I check the "holes" and recreate missing ones. 
It involves https://wiki.mozilla.org/ReleaseEngineering/How_To/Create_AWS_staging_slaves but with different configs (-c config/tst-linux{32,64}) and instance data (-i). Puppet certs are revoked by the wrapper scripts automatically and masters sync the revocation list every 15-30 mins, but this won't block the new slaves to be puppetized. However, this cert workflow will be obsoleted by puppet 3.x.
Flags: needinfo?(rail)
I'm calling this FIXED because there's nothing actionable here, as Rail goes through to fix gaps in slave lists periodically.
Status: NEW → RESOLVED
Last Resolved: 4 years ago
Resolution: --- → FIXED
(Assignee)

Updated

4 years ago
Product: mozilla.org → Release Engineering
For the record, this machine has existed for a long time since it was terminated, but it's been decomm in slavealloc, so it's been started intermittently and has done no work. I un-decomm'ed it slavealloc.
You need to log in before you can comment on or make changes to this bug.