iSCSI Backup Repository

[deduplicat3d]

I currently have a backup job that backs up from two different vcenters. The veeam server resides on vcenter 1 and I have a proxy agent on vcenter 2. The job backs up to a mounted iSCSI volume (windows iscsi initiator) on the veeam server (vcenter 1). I have noticed that the backups for machines running on vcenter 2 have a bottleneck of 99% for "Target" and "Network" and am assuming this is due to the traffic first being sent to the veeam server and then to the actual device (two hops). Is there anyway for the proxy agent to send the traffic directly to the target?

Thanks
Corey

[Gostev]

Your real bottleneck is the target storage speed. Don't pay attention to the network showing 99% - this is a bug in the bottleneck analysis logic. Network counter shows incorrect reading when target is unable to accept data due to being overloaded. This will be fixed in the future updates. Thanks.

[deduplicat3d]

How come the target does not bottleneck when it backs up from the first vcenter? i.e. when the iscsi target only has one hop.

[Gostev]

Well, to comment anything on the performance of the 1st job, I really need to see its bottleneck stats as well... but likely either because of slower source storage throughtput, or due to the different job settings.