Closed
Bug 1370984
Opened 7 years ago
Closed 7 years ago
generic_worker puts hosts directly into service
Categories
(Infrastructure & Operations :: RelOps: Puppet, task)
Infrastructure & Operations
RelOps: Puppet
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: dividehex, Assigned: dividehex)
Details
Attachments
(1 file, 1 obsolete file)
1.13 KB,
patch
|
dustin
:
review+
dividehex
:
checked-in+
|
Details | Diff | Splinter Review |
Including generic_worker module puts hosts directly into service within taskcluster. There is currently no other gate check to prevent that. This means all the mac minis that got imaged in mdc1 a couple weeks ago are running tier 2 jobs and need to be disabled.
Attachment #8875370 -
Flags: review?(wcosta)
Updated•7 years ago
|
Attachment #8875370 -
Flags: review?(wcosta) → review+
Assignee | ||
Comment 1•7 years ago
|
||
That patch is not going to work since it is an agent and not a daemon service in the eyes of launchd. And I can't figure out how to even unload the agent from the cli. Using an exec to pkill it will only cause launchd to relaunch it. So this might mean rm -rf the plist and reboot. :-(
Assignee | ||
Comment 2•7 years ago
|
||
Falling back to removing the plist. We'll reboot the hosts manually with ansible or something to that effect.
Attachment #8875370 -
Attachment is obsolete: true
Attachment #8875474 -
Flags: review?(wcosta)
Updated•7 years ago
|
Attachment #8875474 -
Flags: review?(wcosta) → review+
Assignee | ||
Updated•7 years ago
|
Attachment #8875474 -
Flags: checked-in+
Assignee | ||
Comment 3•7 years ago
|
||
These never got rebooted so they were still taking jobs. Mea cupla. I've since mass rebooted all the testers in mdc1. A few spot checks shows the worker daemon is no longer starting at boot.
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → FIXED
You need to log in
before you can comment on or make changes to this bug.
Description
•