Backup failure/error

[ZeljkoG]

I have VPS server with all of my websites and now i buyed a dedicated server - both servers use Plesk. On VPS i created backup file named pleskbackup using command on ssh: /usr/local/psa/bin/pleskbackup server --skip-logs --output-file=/backup/pleskbackup - file have 2GB and i transfered it via ftp to a new dedicated server.
On dedicated server i used command via ssh:

/usr/local/psa/bin/pleskrestore --create-map pleskbackup -map dump1

and when it`s finished, i used command:

/usr/local/psa/bin/pleskrestore --restore pleskbackup -level clients -map dump1

and in the end, i get this error in pmmcli.log file:

== STDERR ====================
PHP Fatal error: Call to a member function getVendorGuid() on a non-object in /usr/local/psa/admin/plib/backup/Conflicts/Resolver/ResolverPleskModel.php on line 114

7671: 2011-05-28 20:59:53,640 CRITICAL PMMUtility exception:
Subprocess <subprocess[7672] '/usr/local/psa/bin/sw-engine-pleskrun /usr/local/psa/admin/plib/backup/Conflicts/Runner.php --resolve-conflicts --plesk-dump=/usr/local/psa/PMM/rsessions/20110528205953613/dump.xml --owner-guid=b81c4425-20b3-477c-b081-cdf0c162a17f --restore-specification=/usr/local/psa/PMM/rsessions/20110528205953613/restore.xml.conflicted --conflicts-description=/usr/local/psa/PMM/rsessions/20110528205953613/conflict_description --conflicts-resolution-rules=/usr/local/psa/PMM/rsessions/20110528205953613/conflict_resolution_rules --restore-specification-out=/usr/local/psa/PMM/rsessions/20110528205953613/restore.xml --session-path=/usr/local/psa/PMM/rsessions/20110528205953613 --log=conflict-resolve.log'> was finished with exit code 255
== STDOUT ====================

== STDERR ====================
PHP Fatal error: Call to a member function getVendorGuid() on a non-object in /usr/local/psa/admin/plib/backup/Conflicts/Resolver/ResolverPleskModel.php on line 114

Traceback (most recent call last):
File "/usr/local/psa/admin/share/pmmcli/pmmcli.py", line 2106, in main
data_action_response, errcode_response, error_message = actions.get(sys.argv[1][2:])(parameters)
File "/usr/local/psa/admin/share/pmmcli/pmmcli.py", line 1955, in resolve_conflicts
return ActionRunner(ResolveConflictsAction, resolve_conflicts_task_description, None).doActivity()
File "/usr/local/psa/admin/share/pmmcli/pmmcli.py", line 150, in doActivity
result = self.processor.doActivity()
File "/usr/local/psa/admin/share/pmmcli/pmmcli.py", line 1328, in doActivity
rsession.resolveConflictsOnce(rsession.get_plesk_actual_dump(),self.get_resolve_conflicts_task_description().get_conflict_resolution_rules())
File "/usr/local/psa/admin/share/pmmcli/pmmcli_session.py", line 356, in resolveConflictsOnce
self.__doResolveConflictsOnce(plesk_actual_dump, conflict_resolution_rules_object)
File "/usr/local/psa/admin/share/pmmcli/pmmcli_session.py", line 353, in __doResolveConflictsOnce
pmm_conflict_detector.ConflictResolver.run(owner_guid, plesk_actual_dump, stored_restore_specification, self.__restore_specification, self.__conflict_description, stored_conflict_resolution_rules, self.__session_path)
File "/usr/local/psa/admin/share/pmmcli/pmm_conflict_detector.py", line 162, in run
raise PMMUtilityException('Conflict Resolver', x)
PMMUtilityException: was finished with exit code 255

7671: 2011-05-28 20:59:53,640 INFO Outgoing packet:
<?xml version="1.0" encoding="UTF-8"?>
<response>
<errcode>1000</errcode>
<errmsg>pmm utility 'Conflict Resolver' raised an exception. Error code is: 255
See pmmcli.log to find out detailed information on this</errmsg>
</response>


A fast reply would be appreciated.
Thx in front!

 

[comicmaus]

Have you solved the problem in the meantime?

Currently I see the same problem with the same error message on Suse 11.4 with Plesk 10.2.
If I try to migrate a client from one server to another via the Migration Manager, everything is working fine, but if I do a pleskbackup on one side and a pleskrestore on the other, the error appears only a few seconds after starting the restore.

 

[comicmaus]

Igor, where are you ?

The problem still exists. I can't restore clients that have been backed up on Suse 11.3 / Plesk 10.2 to Suse 11.4 / Plesk 10.2. The Migration manager at the webinterface works for the most clients.

 

[IgorG]

Looks like python and libxml2 packages should be reinstalled. But I'm not sure. Detailed investigation directly on server is required.

 

[comicmaus]

Days later, here are just a few thoughts for a solution:

The function getVendorGuid() reads the guid from the psa database. I only found the guid in the clients table. What it reads isn't the guid of the actually restored client, instead it reads the guid of the vendor. When listing all Clients at the Plesk website, this is the contact listed in the Provider column.
So you just have to take care that at the server you are trying to restore, the provider has the same guid as on the server the backup has been made.

There are two ways to do this. First is, to backup the complete server configuration at the old server and restore it at the new machine. This would transfer all guids, also the one from the admin user. (I've not really tested this way!)
Second way is to do some changes in the psa database. I assume that there is no reseller, all clients to be restored are belonging directly to the admin user.
At the original server:
Read the guid of the admin user: SELECT id,guid,vendor_id from clients;
The admin user normally has the id 1 and the vendor_id is NULL.
At the new server:
Change the guid of the existing admin user to the on from old server: UPDATE clients SET guid="<old-server-admin-guid>" where id=1;
If the clients belong to a reseller, do the same steps for the reseller account. Now there shouldn't be a problem with guid any longer. And don't forget to create the map before restoring an clients!

What a nice bug with a completely confusing error message, isn't it?