Please explain botthleneck network

[b.vanhaastrecht]

Hi,

All our backup jobs state botthleneck network. Proxy and repository (local disks) are on the same box. We are back-up thru iSCSI [san], 2x 10GbE, logs confirm this. The NICs are far from being maximum utilized, max of 2Gb/s per NIC. Data is coming from a cluster of Equallogic 10GbE boxes.

All jobs about:
50% source, 40% proxy, 75% network, target 1%

Veeams manual states "Bottleneck - Network — the network queue writer component responsible for getting processed VM data from the backup proxy and sending it over the network to the backup repository or another backup proxy." But in our setup the proxy and repository are on a single box, so I don't understand this. Could someone please explain how the bothlenech network metric is being determend, which performance counters, and what does it mean in a single box setup?

Regards,
Bastiaan

[veremin]

Hi Bastiaan,

Even when both source and target components are deployed on one machine, they still communicate with each other over TCP. The traffic doesn't go outside, however. This communication seems to have been marked as the bottleneck in your case.

Thanks.

[Shestakov]

Bastiaan,
There is no architectures without a bottleneck. One of the components will be a bottleneck anyway.
Having network as a bottleneck in SAN mode means all other components show excellent performance.
Thanks.

[b.vanhaastrecht]

Hi, ok I understand, thank you both for your clarification.

I do have a connected question, but I will create a seperate post for that.