Windows 10 AArch64 opt bitbar workers not taking tasks
Categories
(Taskcluster :: Workers, defect)
Tracking
(Not tracked)
People
(Reporter: apavel, Unassigned)
References
Details
There are jobs in queue for ~464 minute(s)
From IRC: tomprince> (it looks like the workers aren't taking jobs)
Edwin can you take a look?
Comment 1•6 years ago
|
||
Mentioned also in irc, but I suspect either:
- increased windows10-aarch64 load, or;
- some windows10-aarch64 machines have gone offline for an unknown reason
I've noted (as of 2019-07-04 22:45 PDT) that the following machines appear to be offline (not showing in the taskcluster worker listing):
009, 010, 011, 012, 013, 016, 017, 018, 019, 020, 021, 022, 024, 025, 027, 028
I've inquired with Bitbar; as they are based in Pacific timezone it may be another ~8 hours until a response is received.
Reporter | ||
Comment 2•6 years ago
|
||
At this time (10h later) the jobs are still in queue https://irccloud.mozilla.com/file/c862lwGD/image.png
Comment 3•6 years ago
|
||
BItbar worker count is still the same as listed in Comment 1. Bitbar has not yet responded on slack.
Comment 4•6 years ago
|
||
bc was investigating this at one point this week so ni'ing him.
Comment 5•6 years ago
|
||
I don't have access to the projects or devices at bitbar to investigate. I have someone going to the DC this morning to check on devices. I'll ask them to look around but I don't know specifically what to tell them.
Comment 6•6 years ago
|
||
Stanley @ Bitbar is looking at the laptops now. I see them servicing jobs from Friday evening now.
Reporter | ||
Comment 7•6 years ago
|
||
Yes, jobs are green, things look good. Closing for now, please reopen if there are other issues.
Comment hidden (Intermittent Failures Robot) |
Comment 9•6 years ago
|
||
We have fluctuating worker count for windows10-aarch64. According to Bitbar, some of the machines are stuck on login screen (after rebooting post-test execution).
Bitbar suspects that at least some of the machines will have to reimaged/factory reset.
Comment 10•6 years ago
•
|
||
This issue may be resolved by upgrading to generic-worker 15.1.0 (task user management was a bit broken in 14.1.1).
Description
•