Re: [pve-devel] [PATCH qemu-server] api: record VM ID as used after a virtual machine is destroyed

From: Thomas Lamprecht <t.lamprecht@proxmox.com>
To: Severen Redwood <severen.redwood@sitehost.co.nz>,
	pve-devel@lists.proxmox.com
Subject: Re: [pve-devel] [PATCH qemu-server] api: record VM ID as used after a virtual machine is destroyed
Date: Thu, 26 Sep 2024 19:53:03 +0200	[thread overview]
Message-ID: <aa647cfc-14e8-4bb5-b5e1-f3a8dad58070@proxmox.com> (raw)
In-Reply-To: <20240926135516.117065-5-severen.redwood@sitehost.co.nz>

Am 26/09/2024 um 15:52 schrieb Severen Redwood:
> After a virtual machine is destroyed, record that its ID has been used
> via the `PVE::UsedVmidList` module so that the `/cluster/nextids`
> endpoint can later optionally avoid suggesting previously used IDs.
> 
> Co-authored-by: Daniel Krambrock <krambrock@hrz.uni-marburg.de>
> Signed-off-by: Severen Redwood <severen.redwood@sitehost.co.nz>
> ---
>  PVE/API2/Qemu.pm | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/PVE/API2/Qemu.pm b/PVE/API2/Qemu.pm
> index d25a79fe..67a6191f 100644
> --- a/PVE/API2/Qemu.pm
> +++ b/PVE/API2/Qemu.pm
> @@ -2340,6 +2340,7 @@ __PACKAGE__->register_method({
>  	    });
>  	};
>  
> +	PVE::UsedVmidList::add_vmid($vmid);

same here, you write "after a virtual machine is destroyed", but this is quite
a bit before that as the worker might need quite a bit of time to finish, and
that can even fail.

While it's not really causing a issue with recording the VMID as reserved,
that should be evaluated explicitly and mentioned in the commit message.

>  	return $rpcenv->fork_worker('qmdestroy', $vmid, $authuser, $realcmd);
>      }});
>  

_______________________________________________
pve-devel mailing list
pve-devel@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel