Standalone backup agent for Microsoft Windows servers and workstations (formerly Veeam Endpoint Backup FREE)
Post Reply
dcampregher
Service Provider
Posts: 135
Liked: 6 times
Joined: Mar 23, 2016 5:57 pm
Full Name: Diogo Campregher
Contact:

High data read after get on dedup volume

Post by dcampregher » Sep 13, 2019 2:39 am

Hello!
A customer with a big Windows Cluster Fileserver ( 180 TB ) with Veeam Agent for Windows run the backups from the january of this year until 03 weeks ago with no problems and was all right until them resolve to activate the Deduplication on the volumes. After do this, the backups that worked for fewer hours ( 03 or 04 hours ) now run for all the day ( 24 hours ) and in some days go the the next day, overriding the backup window.

I note in the statistcs that the Data Read now is very big in contrast with the Data Transferred, so even with the very high performance of the cluster, there is no time to conclude the backups in the required Window.

Anyone know this behavior and how workaround ( in the Windows or the Veeam Side ) ?

ejenner
Expert
Posts: 377
Liked: 55 times
Joined: Mar 23, 2018 4:43 pm
Full Name: EJ
Location: London
Contact:

Re: High data read after get on dedup volume

Post by ejenner » Sep 13, 2019 8:31 am

I'm not sure I fully understand what you're describing but I know from our own experience that a possible hazard with fileserver clusters and Windows Agent is the possibility for the backup to start again from the beginning. Might that have happened in this case? For us a STOP error would cause a resynchronization of our DFS fileserver and that made Veeam think the files were all new and had to be backed up again. The workaround was to target the active cluster rather than the passive synchronized copy. In our setup it's always the passive copy which resynchronizes, the active copy is treated as the master which changes are copied from so much more stable and a better source for backing up.

The other thing we've seen is that on larger file server backups the healthcheck was stopping the backups from completing. They would actually stall at 99% and run for days at that stage without ever getting to 100%. This was happening somewhere between 2 and 6TB. So for the larger jobs we turned off the healthcheck and the backups then completed.

dcampregher
Service Provider
Posts: 135
Liked: 6 times
Joined: Mar 23, 2016 5:57 pm
Full Name: Diogo Campregher
Contact:

Re: High data read after get on dedup volume

Post by dcampregher » Sep 13, 2019 11:00 am

Hello! I understand the case where the files need to be all read again but in my case this is not happening.
Simply the Veeam now is reading a lot of 'data' from the volumes but without sending it to the repository. I think that the Windows Dedup services is changing blocks on the volumes, doing that the Veeam read this blocks. Or not?

ejenner
Expert
Posts: 377
Liked: 55 times
Joined: Mar 23, 2018 4:43 pm
Full Name: EJ
Location: London
Contact:

Re: High data read after get on dedup volume

Post by ejenner » Sep 16, 2019 10:11 am

Windows 2016 has built-in tools in the performance management / task manager area which are enabled by default. You can open the task manager and change the view until you can see what is using your disk. You have to open 'Resource Manager' and move to the disk tab. Then expand the different sections below and look for disk read/writes

Post Reply

Who is online

Users browsing this forum: Bing [Bot] and 3 guests