QA (public) » openQA Project (public) » openQA Infrastructure (public)

Julie_CAO wrote in #note-3:

Is this issue related to https://progress.opensuse.org/issues/177892?

When you link tickets please use the format #<id> so that we have direct preview of the subject and status of the linked ticket.

No, this is not related to #177892. If the worker instances would be unavailable due to limited number of available connection slots then the word "limited" would show up in the status column.

when shall we expect them back to the pool? Our tests are blocked by missing these 2 machines.

oh, ok. You created the ticket with priority "normal" so I assumed that this was not immediately blocking. Feel welcome to use "High" for such cases in the future. We should work on #178015 first which is why we didn't notice that those worker instances are offline when they shouldn't be

Those units are masked:

martchus@worker33:~> sudo systemctl status openqa-worker-auto-restart@{16,17}.service 
Warning: The unit file, source configuration file or drop-ins of openqa-worker-auto-restart@16.service changed on disk. Run 'systemctl daemon-reload' to reload units.
Warning: The unit file, source configuration file or drop-ins of openqa-worker-auto-restart@17.service changed on disk. Run 'systemctl daemon-reload' to reload units.
○ openqa-worker-auto-restart@16.service
     Loaded: masked (Reason: Unit openqa-worker-auto-restart@16.service is masked.)
    Drop-In: /etc/systemd/system/openqa-worker-auto-restart@.service.d
             └─20-nvme-autoformat.conf, 30-openqa-max-inactive-caching-downloads.conf
     Active: inactive (dead)

○ openqa-worker-auto-restart@17.service
     Loaded: masked (Reason: Unit openqa-worker-auto-restart@17.service is masked.)
    Drop-In: /etc/systemd/system/openqa-worker-auto-restart@.service.d
             └─20-nvme-autoformat.conf, 30-openqa-max-inactive-caching-downloads.conf
     Active: inactive (dead)

That means someone intentionally disabled them. The question is what the intention was.

Judging by the job history on https://openqa.suse.de/admin/workers/2662 and https://openqa.suse.de/admin/workers/2663 those worker slots don't seem to be broken. So I took the liberty of umasking them.

Note that many other units are masked on this host as well and I don't know why. I will not change that because there probably was and maybe still is a reason for that.

EDIT: I now unmasked everything because the masking wasn't done consistently leading to failed systemd units.

Setting due date based on mean cycle time of SUSE QE Tools

Not sure why those worker slots were disabled, though.

I guess noone remembers why that worker was left in that state. So I'm just considering this resolved now.

Note that I have now created one failing unit on purpose for another ticket (see #178015#note-8) but it will not interfere with any production worker slots.