I have a Vista client which in the past needed a maximum of an hour for Indexing and afterwards executed Backup. Now after 24 hours of Indexing I have stopped it.
In between have reinstalled client, deleted all backups for this client in server, rebooted client and server after deletion and reinstallation, deleted all old shadow copies and still no change. Imagecopy works, client link to server works, no messages in protocol files client or server. When tracing client I find UrBackupClientBackend.exe reading through multiple shadowcopyfiles (with very low speed, 4k/sec), writing into BackupClientDBJournal and various tempfiles, but obviouisly never finishing. Increasing Priority of BackendClient doesn’t change anything, machine is idle.
Any idea what may cause this behaviour, or how to identify what the issue is ?
I think that might be the correct direction. Unfortunately changing the priority of the whole process does not help. Can you use ProcessHacker http://processhacker.sourceforge.net/downloads.php to change the priority of the indexing thread, especially the IO-priority?
- Don’t forget to run process hacker as administrator, otherwise it won’t show the priorities
- Click on UrBackupClientBackend.exe
- Click on the threads tab
- Increase the IO-priority of the yellow -4 thread
In my case the indexing became really slow when the software raid was being rebuild. Unfortunately one cannot fine tune the IO-priority much. I can only set it to background or not.