[rear-users] Failed Restore

Gary Schiltz gary at cloudforestsystems.com
Wed Apr 16 23:08:59 CEST 2014


Arggh, when reading the above link from the pipermail archive, it
includes the period at the end of my sentence. The correct link is
https://github.com/rear/rear/issues/390 which should not have a period
at the end.

On Wed, Apr 16, 2014 at 4:01 PM, Gary Schiltz
<gary at cloudforestsystems.com> wrote:
> I got the idea that the problem may have been a failing disk drive, so
> just for grins I replaced it, with the same result. I went ahead and
> opened issue https://github.com/rear/rear/issues/390.
>
> On Wed, Apr 16, 2014 at 11:09 AM, Gary Schiltz
> <gary at cloudforestsystems.com> wrote:
>> I will open an issue as you recommended, but I will try a few things
>> first and report back. The interesting thing is that it appears that
>> LVM may already be configured (at least there are two files in
>> /dev/mapper, i.e. the ones that represent my two logical volumes) when
>> the script is trying to create the RAID device, and LVM in my config
>> uses a logical volume that is provided by the RAID device. In fact, if
>> I abort the installation and cat /proc/mdstat, it shows the RAID
>> device part way though syncing. Perhaps the OS has detected the
>> existing layout and lvm and/or mdadm have started and are holding on
>> to the drives?
>>
>> On Wed, Apr 16, 2014 at 6:26 AM, Gratien D'haese <gratien.dhaese at it3.be> wrote:
>>> Hi,
>>> I think it would be better to create an issue for this problem as I cannot
>>> tell by heart what caused it.
>>> Could you also add the disklayout.conf file and perhaps the diskrestore.sh
>>> script created by rear recover.
>>> My best guess is that you cannot stop the md1 mirroring during is setup
>>> process due to a lock created by mdadm?
>>>
>>> best regards,
>>> Gratien
>>>
>>> On Tue, 15 Apr 2014 10:31:42 -0500, Gary Schiltz
>>> <gary at cloudforestsystems.com> wrote:
>>>> (Apologies if this appears twice; I posted it yesterday from a
>>>> different email address that is not a member of the list)
>>>>
>>>> Today I tried to use ReaR to do a bare metal restore of a new server
>>>> that I'm setting up. To summarize, it failed partway through the "rear
>>>> recover" phase, apparently during trying to stop a software RAID
>>>> device. The server had Ubuntu 12.04.4 and the following describes the
>>>> disk/partitioning setup.
>>>>
>>>> Three physical disks
>>>>   /dev/sda    1 TB SATA, msdos partition map, bootable flag set, GRUB
>>>>   installed
>>>>   /dev/sda1   200 MB partition for software RAID1  (for /dev/md0)
>>>>   /dev/sda2   Remainder of disk for software RAID1 (for /dev/md1)
>>>>
>>>>   /dev/sdb    1 TB SATA, msdos partition map, bootable flag set?, GRUB
>>>>   installed
>>>>   /dev/sdb1   200 MB partition for software RAID1  (for /dev/md0)
>>>>   /dev/sdb2   Remainder of disk for software RAID1 (for /dev/md1)
>>>>
>>>>   /dev/sdc    1 TB SATA, guid partition map
>>>>     No RAID or LVM, single ext4 /dev/sdc1, mount point /backup
>>>>
>>>> Two software RAID devices
>>>>   /dev/md0  RAID1, uses /dev/sda1 & /dev/sdb1, mount point /boot
>>>>   /dev/md1  RAID1, uses /dev/sda2 & /dev/sdb2, physical volume for LVM
>>>>
>>>> LVM
>>>>   Volume group named system-vg, uses /dev/md1 RAID device
>>>>   Logical volume system--vg-root--lv 100 GB ext4, mount point /
>>>>   Logical volume system--vg-data--lv 900 GB ext4, mount point /data
>>>>
>>>> My /etc/rear/local.conf is as follows:
>>>>   OUTPUT=ISO
>>>>   BACKUP=NETFS
>>>>   BACKUP_URL=file:///backup/relax-and-recover
>>>>   EXCLUDE_MOUNTPOINTS=( '/data' )
>>>>
>>>> I exclude /data since I will back it up separately, and /backup is
>>>> automatically excluded since ReaR uses it to write the ISO and
>>>> tarball.
>>>>
>>>> I created the backup using "rear -v mkbackup", then burned a CD from
>>>> the resulting ISO file. When I boot from the CD, everything seems okay
>>>> in the boot. Just before typing "rear recover", I do "mkdir /backup"
>>>> and "mount /dev/sdc1 /backup", since this is where mkbackup wrote to.
>>>> ReaR seems to get through the process of partitioning the disks, but
>>>> eventually fails with the following in the log:
>>>>
>>>> +++ Print 'Creating software RAID /dev/md1'
>>>> +++ test 1
>>>> +++ echo -e 'Creating software RAID /dev/md1'
>>>> +++ test -b /dev/md1
>>>> +++ mdadm --stop /dev/md1
>>>> mdadm: Cannot get exclusive access to /dev/md1:Perhaps a running
>>>> process, mounted filesystem or active volume group?
>>>> 2014-04-14 15:55:49 An error occured during layout recreation.
>>>>
>>>> Any ideas? I'd be glad to supply the rest of the log, the contents of
>>>> /var/lib/rear, etc... Thanks in advance.
>>>> _______________________________________________
>>>> rear-users mailing list
>>>> rear-users at lists.relax-and-recover.org
>>>> http://pikachu.3ti.be/mailman/listinfo/rear-users
>>>
>>> --
>>> Gratien D'haese
>>> IT3 Consultants bvba
>>> Vennestraat 15, B-2560 Nijlen
>>> _______________________________________________
>>> rear-users mailing list
>>> rear-users at lists.relax-and-recover.org
>>> http://pikachu.3ti.be/mailman/listinfo/rear-users


More information about the rear-users mailing list