[pve-devel] [PATCH ha-manager 00/18] HA rules fixes + cleanup + performance improvements

[Daniel Kral]

Thanks a lot for the review and testing! Appreciate it!

On Fri Aug 29, 2025 at 2:44 PM CEST, Michael Köppl wrote:
> Tested this series by:
> - Creating rules with resources that were set to ignored, checked if the
>   resources are not shown as dependent resources when migrating

One thing that I want to point out here is that 'ignored' HA resources
are still accounted for when checking the feasibility and transforming
HA rules as that state is not really defined yet.

The 'ignored' HA resources state makes them be ignored for the HA
Manager, but I wondered how we should handle that for the HA rules
config, checks, etc.

The reason currently is that putting HA resources from any state to
'ignored' and vice-versa could make HA rules infeasible or feasible
again. Especially the former case (making HA rules infeasible through
changing HA resource states) didn't fit right for me as it feels weird
that changing HA resource states could do that.

I'll check that again when sending patches for #6613, but I'm open for
thoughts about that behavior.

> - Also tested what happens when the resource state was recently changed.
>   Noticed that when changing from 'started' to 'ignored', it can take a
>   few seconds before the UI reflects that (e.g. when using the migrate
>   dialog). This isn't unexpected, but wanted to note it nonetheless.

What do you mean there exactly? Do you mean the state that is shown in
the HA Resources panel or other places in the UI?

> - Checking that conflicts between positive resource affinity rules and
>   node affinity rules are now detected correctly that weren't detected
>   before
> - Manually created various combinations of node affinity and resource
>   affinity rules and checked that they are applied
> - Also re-ran some group-to-rules migrations, just to be sure
>
> I did not notice any problems apart from the minor delay in the UI noted
> above. The rules are more robust now and the changes solve some problems
> that I had encountered before.
>
> I also had a closer look at the code and left some comments on the
> individual patches, but nothing major. I had a particularly close look
> at the compiled rules to be certain that they produce the expected
> structure.
>
> With my comments on the individual patches addressed, please consider
> this series
>
> Reviewed-by: Michael Köppl <m.koeppl at proxmox.com>
> Tested-by: Michael Köppl <m.koeppl at proxmox.com>