Volume restore fails with recovery media for CentOS

[LaHo]

Hello,
I did generade a volume based backup for a virual CentOS 6.2.
Restoring single files from the backup without the recovery iso is working.
Trying to restore the complete volume with the provided recovery ISO (veeam-recovery-media-1.0.0.499.iso) ends in error

By selecting the restore point the error:CMcTableWidget: Unable to select row #[18446744073709554615] has been reported.


If you need further information please let me know.

[PTide]

Hi,

Please upload your Recovery Media logs (/usr/share/Veeam/LPB on recovery media) as well as backup logs (/var/log/veeam on the original machine) to the dropbox (or to some other place of your choice) and post the link.

Thanks

[LaHo]

Hello,

Meanwhile I created a new virtual machine on the esx. The recovery media iso is able to restore the backup on that new virtual machine. The difference I see is that the failing VM has Hardware Version 8 while the new VM has Hardware Version 10. After upgrading to Hardware Version 10 the restore still ends in the same error. Also I did clear the partition table so there is an empty HDD. The difference is now only in the Virtual Hardware maybe.

The workaround to be able to restore was to generate a new VM. The backup file is functional but why is it failing on the origin VM? ISO is the same, VM NIC is the same.

Here are the two logfiles files


Thanks for your help

[PTide]

One more question - did you perform system upgrade after veeam installation? It can be seen from logs that the kernel version has changed from 2.6.32-504.1.3.el6.x86_64 to 2.6.32-642.1.1.el6.x86_64. Could you please upload the whole /var/log/veeam directory?

Thanks!

[LaHo]

Yes I did the kernel upgrade.
First I installed the Veeamsnap driver.
Second I tried to install the Software package.
Then I did update the kernel to match kernel-devel and the load of veamsnap was possible.
Afterall I installed the Veeam rpm.

The /var/log/veeam directory is uploaded to the other files.

[LaHo]

I located the problem.

The problem is related to the existing partition table.
My first try to delete the partitions with fdisk did not fix it. Also to writing a new DOS partition table fdisk (o) does not solve the error.
Only by providing an empty harddisk or overwriting the first MB to clear all partition information makes the selection of the restore point and restore possible.

If you want I can upload a 'dd' backup of my first MB of my harddisk to the cloud share and you can try it on your own.

[PTide]

Our QA team has already reproduced that behaviour, but thanks for the assistance!