Re: [Clonezilla-live] Bug found in lvm restoration, patch for this & test new lives CD

[Steven S. <st...@nc...>]

Hi Gerald,
That's great. Thanks for this bug report. We will patch this in the next
release. When a newer one is released, I will let you know and please
give it a try then.

Thanks again.

Regards,
Steven.

Gerald HERMANT wrote:
> Hello Steven,
> 
> Thanks for your very good job in Clonezilla.
> 
> I find a pb in lvm restoration in some case.
> The problem:
>  - Restoration process seems good, but kernel can't find his root (in LVM).
> 
> Analyse :
>  - pvdisplay => PV is created
>  - vgdisplay => No VG is created
>    My source data (/home/partimage) are remote on a samba server.
>    in ocs-functions (and ocs-onthefly) the section is
> 
>    # create PV first
>    echo "Creating the PV... "
>    while read vg dev uuid; do
>     pvcreate -ff --yes --uuid $uuid $dev
>    done < $PV_PARSE_CONF
>    echo "done!"
>   
>    # Restore the vg conf
>    echo "Restoring the VG config... "
>    while read vg dev uuid; do
>     vgcfgrestore -f *$target_dir/lvm_$vg.conf* $vg 2>/dev/null
>    done < $PV_PARSE_CONF
>    echo "done!"
>    #
> 
> I notice 2 things :
> 
>     * vgcfgrestor -f $target_dir/lvm_$vg.conf $vg => Use system function
>       mmap for read the file, and this mmap function is not available on
>       every remote filesystem (not in samba in my case). Then because of
>       this call failed, and restoration of vg failed.
>     * It could be better to use : pvcreate -ff --yes --uuid $uuid
>       ---restorefile /tmp/lvm_$vg.conf $dev, because, pvcreate in this
>       case, create pv much identical than without (same topology if
>       possible - see man pvcreate). pvcreate use mmap syustem call to.
> 
> Patchs :
> --- ocs-functions-ORIG    2008-02-14 08:50:36.732643764 +0000
> +++ ocs-functions    2008-02-14 09:42:13.013189303 +0000
> @@ -3307,14 +3307,20 @@
>    # create PV first
>    echo "Creating the PV... "
>    while read vg dev uuid; do
> -    pvcreate -ff --yes --uuid $uuid $dev
> +    cp $target_dir/lvm_$vg.conf /tmp/  # mmap not available on remote disk
> +    pvcreate -ff --yes --uuid $uuid --zero y --restorefile
> /tmp/lvm_$vg.conf $dev
> +    rm /tmp/lvm_$vg.conf
>    done < $PV_PARSE_CONF
>    echo "done!"
>   
>    # Restore the vg conf
>    echo "Restoring the VG config... "
>    while read vg dev uuid; do
> -    vgcfgrestore -f $target_dir/lvm_$vg.conf $vg 2>/dev/null
> +    cp $target_dir/lvm_$vg.conf /tmp  # mmap failed on some network fs
> +    vgcfgrestore -f /tmp/lvm_$vg.conf $vg 2>/dev/null
> +    rm /tmp/lvm_$vg.conf
>    done < $PV_PARSE_CONF
>    echo "done!"
>    #
> 
> Same patch should be apply to ocs-onthefly.
> With this, all things is OK. I only copy localy file before starting
> commands, and after remote temporary file.
> Don't you think it could be good to add command exit code for stop
> restoration if pre-requires are not done ? [I know it could be difficult
> in bash, but...]
>  
> 
> I test both new .iso images :
>  -clonezilla-live-1.0.9-8.iso => kernel is to old, and my nic carte
> e1000 (last model in HP desktop), is not recognize.
>  -clonezilla-live-20080211-hardy-experimental.iso => Work well, and my
> e1000, is well seen.
> 
> Thanks.
> 
> Gerald
> 
> -- 
> ASTREL "Raison & Télécom"
> ghe...@as... - http://www.astrel.fr
> 


-- 
Steven Shiau <steven _at_ nchc org tw> <steven _at_ stevenshiau org>
National Center for High-performance Computing, Taiwan.
http://www.nchc.org.tw
Public Key Server PGP Key ID: 1024D/9762755A
Fingerprint: A2A1 08B7 C22C 3D06 34DB  F4BC 08B3 E3D7 9762 755A