resalloc: Unknown pool

[xsuchy]

In the resalloc log I see:

ERROR 2023-11-03 15:53:13,939 PID:316334:139871599429440(Manager) Unknown pool aws_x86_64_powerful_normal_prod to cleanup
ERROR 2023-11-03 15:53:13,948 PID:316334:139871599429440(Manager) Unknown pool aws_aarch64_powerful_normal_prod to cleanup

Not sure if this is Copr's miscongiruation or resalloc issue.

[praiskup]

Seems like aws_x86_64_powerful_spot_prod vs. aws_x86_64_powerful_normal_prod problem. The "normal" seems new in "powerful" category. This is related to a hotfix for praiskup/resalloc#128 - but it seems obvious we should have a better error message.

[praiskup]

Btw., aws_x86_64_powerful_normal_prod is not going to allocate resources (normally), since it has lower priority compared to aws_x86_64_powerful_spot_prod. Even if EC2 refuses to reserve spot instances for us, resalloc will continue asking for them never falling back to the "normal" category.

Normal high-performance builders will only be allocated if all the spot builders are already taken (or at least requested and being started).

[praiskup]

Soft-blocked on this: https://pagure.io/fedora-infrastructure/issue/11602 I'll continue tomorrow.

[praiskup]

New resalloc deployed to production.