[pve-devel] [PATCH v2 container 2/2] clear /etc/machine-id also after container clone

[Fabian Grünbichler]

On May 26, 2021 4:18 pm, Oguz Bektas wrote:
> pass $clone=1 to avoid removing the file. instead we truncate it to an
> empty file
> 
> Signed-off-by: Oguz Bektas <o.bektas at proxmox.com>
> ---
> 
>  src/PVE/API2/LXC.pm | 19 ++++++++++++++++++-
>  1 file changed, 18 insertions(+), 1 deletion(-)
> 
> diff --git a/src/PVE/API2/LXC.pm b/src/PVE/API2/LXC.pm
> index a9ea3a6..413f466 100644
> --- a/src/PVE/API2/LXC.pm
> +++ b/src/PVE/API2/LXC.pm
> @@ -1594,7 +1594,24 @@ __PACKAGE__->register_method({
>  	};
>  
>  	PVE::Firewall::clone_vmfw_conf($vmid, $newid);
> -	return $rpcenv->fork_worker('vzclone', $vmid, $authuser, $realcmd);
> +
> +	my $task = eval {
> +	    return $rpcenv->fork_worker('vzclone', $vmid, $authuser, $realcmd);

this forks the task worker that does the actual cloning of 
mountpoints/volumes, so after this point that code ($realcmd) will run 
concurrent to the rest of the API worker itself handling the request.

doing anything after forking the task worker is almost always wrong.

> +	};
> +	if (my $err = $@) {
> +	    warn $@ if $@;
> +	    die $err;
> +	}
> +
> +	my $lastconf = PVE::LXC::Config->load_config($newid);

this config might or might not contain any of the updated/cloned 
volumes, this is entirely up to the speed of cloning

> +	my $rootdir = PVE::LXC::mount_all($newid, $storecfg, $lastconf, 1);

so this might or might not mount anything?

> +	my $lxc_setup = PVE::LXC::Setup->new($lastconf, $rootdir);
> +	$lxc_setup->clear_machine_id($lastconf, 1);
> +	PVE::LXC::umount_all($newid, $storecfg, $lastconf, 1);

in which case this might or might not do anything

> +
> +	return $task;

why do you not simply clear the machine ID at the end of the task 
worker?

> +
> +
>      }});
>  
>  
> -- 
> 2.20.1
> 
> 
> _______________________________________________
> pve-devel mailing list
> pve-devel at lists.proxmox.com
> https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
> 
> 
>