DirectNFS on Nutanix

[lowlander]

Hi,

there is a great whitepaper for Nutanix and Veeam for version 9.5. My question is regarding the implementation of directNFS on v10.

We see great read performance running back-up jobs using on-host proxy servers on ESXi hosts on nutanix nodes. A node has two 10 Gbps network adapters and we are running a virtual repository also connected to this 10 Gbps network.

While the read performance is using 10 Gbps, it seems to me that write performance from proxy to the repository is capped by 1 Gbps. What speed should be expect from a full backup using directNFS ?

[Andreas Neufert]

Can you check your write speed within a regular VM with a test tool. Should be same 1Gbps speed ?

[PetrM]

Hello,

One more idea is to check "bottleneck" in the job statistics window: you will see "Target" if the job performance is limited by write speed to the repository or "Network" if the data transfer from proxy to repository is the slowest operation.

You can test repository performance by using this tool or you can test network channel by this one depending on the "bottleneck" reported by job.

Thanks!

[lowlander]

spot on ! target server was the bottleneck. We have tested with the speed tool. Storage on the repo was the bottlneck