[pve-devel] [PATCH v2 qemu-server 1/1] Do not start VM twice when rollbacking with --start

[Fabian Grünbichler]

On November 21, 2022 2:13 pm, Stefan Hanreich wrote:
> When rollbacking to the snapshot of a VM that includes RAM, the VM
> gets started by the rollback task anyway, so no additional start task is
> needed. Previously, when running rollback with the --start parameter
> and the VM snapshot includes RAM, a start task was created. That task
> failed because the VM had already been started by the rollback task.
> 
> Additionally documented this behaviour in the description of the --start
> parameter
> 
> Signed-off-by: Stefan Hanreich <s.hanreich at proxmox.com>
> ---
> 
> Changes v1 -> v2:
> Do not parse config for checking type of snapshot but rather directly check
> whether VM is running or not via check_running()
> 
>  PVE/API2/Qemu.pm | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/PVE/API2/Qemu.pm b/PVE/API2/Qemu.pm
> index 6bdcce2..691202d 100644
> --- a/PVE/API2/Qemu.pm
> +++ b/PVE/API2/Qemu.pm
> @@ -5064,7 +5064,8 @@ __PACKAGE__->register_method({
>  	    snapname => get_standard_option('pve-snapshot-name'),
>  	    start => {
>  		type => 'boolean',
> -		description => "Whether the VM should get started after rolling back successfully",
> +		description => "Whether the VM should get started after rolling back successfully."
> +		    . " A VM will always be started when rollbacking a snapshot with RAM included, regardless of this parameter.",

this is worded a bit weird (I don't think that "rollbacking" is a word ;)), how about:

. "(Note: VMs will be automatically started if the snapshot includes RAM.)",

>  		optional => 1,
>  		default => 0,
>  	    },
> @@ -5091,7 +5092,7 @@ __PACKAGE__->register_method({
>  	    PVE::Cluster::log_msg('info', $authuser, "rollback snapshot VM $vmid: $snapname");
>  	    PVE::QemuConfig->snapshot_rollback($vmid, $snapname);
>  
> -	    if ($param->{start}) {
> +	    if ($param->{start} && !PVE::QemuServer::check_running($vmid)) {

unless I am missing something, this should use

 PVE::QemuServer::Helpers::vm_running_locally($vmid)

we are holding the guest migration lock for the whole rollback worker, and
snapshot_rollback loads the config, so we know it is on the current node at this
point and just checking whether a matching qemu process is running after the
rollback is enough.

>  		PVE::API2::Qemu->vm_start({ vmid => $vmid, node => $node });
>  	    }
>  	};
> -- 
> 2.30.2
> 
> 
> _______________________________________________
> pve-devel mailing list
> pve-devel at lists.proxmox.com
> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
> 
> 
>