[pve-devel] [PATCH container v2 5/6] migration: only migrate volumes used by the guest

Fiona Ebner f.ebner at proxmox.com
Mon May 22 17:00:21 CEST 2023


Am 12.05.23 um 14:40 schrieb Aaron Lauterer:
> When scanning all configured storages for volumes belonging to the
> container, the migration could easily fail if a storage is not
> available, but enabled. That storage might not even be used by the
> container at all.
> 
> By not doing that and only looking at the disk images referenced in the
> config, we can avoid that.
> We need to add additional steps for pending volumes with checks if they
> actually exist. Changing an existing mountpoint to a new volume
> will only create the volume on the next start of the container.
> 
> The big change regarding behavior is that volumes not referenced in the
> container config will be ignored.  They are already orphans that used to
> be migrated as well, but are now left where they are.
> 
> Signed-off-by: Aaron Lauterer <a.lauterer at proxmox.com>
> ---
>  src/PVE/LXC/Migrate.pm | 47 +++++++++++++++---------------------------
>  1 file changed, 17 insertions(+), 30 deletions(-)
> 
> diff --git a/src/PVE/LXC/Migrate.pm b/src/PVE/LXC/Migrate.pm
> index ccf5157..8e11be7 100644
> --- a/src/PVE/LXC/Migrate.pm
> +++ b/src/PVE/LXC/Migrate.pm
> @@ -195,6 +195,17 @@ sub phase1 {
>  
>  	return if !$volid;
>  
> +	# check if volume exists, might be a pending new one
> +	eval {
> +	    PVE::Storage::path($self->{storecfg}, $volid);
> +	};
> +	if ($@ =~ m/^unable to parse/) {

I'd really like to avoid such manual error message matching, especially
across package boundaries. PVE::Storage::path() calls parse_volume_id(),
but that also works for storeid:1 for example, so not triggering an
error for such pending changes. PVE::Storage::path() then calls the
plugin's path(), but the plugin can use whatever error message it likes,
so matching for something specific only covers certain plugins. Can't we
just filter out the yet-to-be-created volumes by matching the NEW_DISK_RE?

> +	    $self->log('info', "volume '$volid' does not exist (pending change?)");
> +	    return;
> +	} elsif ($@) {
> +	    die $@;
> +	}
> +
>  	my ($sid, $volname) = PVE::Storage::parse_volume_id($volid);
>  
>  	# check if storage is available on source node

(...)

> +    # first all volumes referenced in snapshots
>      foreach my $snapname (keys %{$conf->{snapshots}}) {
>  	&$test_volid($conf->{snapshots}->{$snapname}->{'vmstate'}, 0, undef)
>  	    if defined($conf->{snapshots}->{$snapname}->{'vmstate'});
>  	PVE::LXC::Config->foreach_volume($conf->{snapshots}->{$snapname}, $test_mp, $snapname);
>      }
>  
> +    # then all pending volumes
> +    if (defined $conf->{pending} && %{$conf->{pending}}) {

Style nit: please use parentheses for defined

> +	PVE::LXC::Config->foreach_volume($conf->{pending}, $test_mp);
> +    }
> +
>      # finally all current volumes
>      PVE::LXC::Config->foreach_volume_full($conf, { include_unused => 1 }, $test_mp);
>  





More information about the pve-devel mailing list