Availability for the Always-On Enterprise
Post Reply
davemann
Novice
Posts: 5
Liked: 2 times
Joined: Jul 15, 2016 8:40 pm
Full Name: Dave Reichert
Contact:

Parallelism in Restores/Backup Copies? (HPE StoreOnce)

Post by davemann » Jun 11, 2018 4:44 pm

Hey there -

I was just wondering... Why is Veeam unable to read from multiple disk sources of a single VM while restoring, or during backup copies? It seems to be able to write to multiple files/volumes just fine during a backup job (for any given VM).

I have some fairly large VM's that have multiple volumes, and my primary backup device is a dedupe appliance (HPE StoreOnce), and this is a major drawback when restoring large VM's with multiple volumes, as well as with Backup Copy Jobs. I'm struggling to restore/copy single large VM's within a reasonable amount of time, and in one particular case, it's practically impossible.

When a job has multiple VM's, Veeam seems to have no issue reading from at least one volume/file from each VM's backup, which substantially improves the overall restore speed of all VM's in the job, but my issue is not with performance on a large number of VM's, just with the speed of many large volumes that are part of one VM.

HPE StoreOnce performs best when multiple streams are established for read/write, but it doesn't seem I'm able to take advantage of this using Veeam in this scenario, which is a bummer, and starting to become an operational issue.

davemann
Novice
Posts: 5
Liked: 2 times
Joined: Jul 15, 2016 8:40 pm
Full Name: Dave Reichert
Contact:

Re: Parallelism in Restores/Backup Copies? (HPE StoreOnce)

Post by davemann » Jun 11, 2018 6:19 pm

FWIW - I am using Catalyst over IP, and utilize the "Architecture B" configuration from the configuration guide, with my remote Catalyst repository using the "Gateway server and StoreOnce are connected over WAN" setting, as I am leveraging Catalyst source-side dedupe for my copy jobs.

Post Reply

Who is online

Users browsing this forum: Baidu [Spider] and 38 guests