Re: [RFC ha-manager 16/21] sim: hardware: add set-dynamic-stats for services

["Dominik Rusovac" <d.rusovac@proxmox.com>]

On Wed Mar 18, 2026 at 5:54 PM CET, Thomas Lamprecht wrote:
> On Tue, 17 Feb 2026, Daniel Kral wrote:
>> --- a/src/PVE/HA/Sim/Hardware.pm
>> +++ b/src/PVE/HA/Sim/Hardware.pm
>> @@ -213,6 +216,25 @@ sub set_static_service_stats {
>> [...]
>> +    if (my $memory = $new_stats->{mem}) {
>> +        $stats->{$sid}->{mem} = $memory;
>> +    }
>> +
>> +    if (my $cpu = $new_stats->{cpu}) {
>
> Note that above will skip a value of 0 or 0.0 since those are falsy.
> Setting a service's CPU to 0.0 (idle) can be valid though?
> If, rather check with if (defined(my $cpu = ...)) { instead.
> Similar issue with `die ... if !$val` in the command handler.

Good catch, thx! Indeed 0 or 0.0 should be valid.

>
>> +        die "overcommitted cpu on '$sid'"
>> +            if $stats->{$sid}->{cpu} > $stats->{$sid}->{maxcpu};
>
> The overcommit dies are quite strict for a simulator and AFAICT prevent
> testing overcommitted scenarios, which are common in real clusters.

Eventually, instead of dying, adding a log on warn for this is a better idea?