Synology RS3614XS replacement for Drobo 1200i

[jeffmetcalf]

I was just wondering if anyone was using for a Synology RS3614XS as a repository for Veeam B&R and how it performed for you. We're currently using a Drobo 1200i and we're finding out that it's just too slow to handle the amount of data we're backing up.

Thanks

[foggy]

Aside from the particular devices, could you please describe what kind of performance issues do you have and how do you add the NAS device to Veeam B&R as a repository? What backup mode do you use (reversed or forward incremental)? Probably we could give some general recommendations on optimizing backup performance in your scenario.

Also, there's a big topic regarding recommended backup targets, worth reviewing: recommendations for backup storage, backup target

[jeffmetcalf]

The Drobo is connected via an iSCSI volume to a physical windows server. We use reverse incremental for all backups. We use Veeam B&R for backup, and Zerto for replication, so we move a lot of bits all of the time. We're not really having trouble with Veeam or the Drobo, but the bottleneck is causing latency issues on the rest of the storage, between Veeam, Zerto, and the SQL T-Log backups that we're running.

Max iops for the drobo is 500-600 for the configuration we have (no SSD). Our backup bottleneck is ALWAYS target, and we observed outbound IOPS from our storage (PS6100s or MD3220s) at 700 and higher.

[foggy]

Having target as a bottleneck is typical for reversed incremental mode, since it produces more I/O load on the target storage. Switching to forward incremental should increase the job processing rate.

[jeffmetcalf]

I know. I'm willing to live with that for the gain in restore time.

[Vitaliy S.]

Hi Jeff,

Do you observe a significant difference in restore times for jobs using different backup modes? What is your current retention policy setting?

Thanks!

[jeffmetcalf]

Thankfully I've only had to go thru one actual restore, so I have no benchmarks to compare to. I keep only 5-10 restore points, depending on the server.

[veremin]

I'm willing to live with that for the gain in restore time.

Could elaborate on it? I'm asking , since with different backup modes there should not be any meaningful difference between time of their restore operations; especially, with such short incremental chains. Or you saw the opposite?

Thanks.

[Vitaliy S.]

Thankfully I've only had to go thru one actual restore, so I have no benchmarks to compare to. I keep only 5-10 restore points, depending on the server.

Given these settings job restore performance should be the same, no matter what backup mode you use.