shared NFS performance issue

[Roshan]

Target: Centos 7.8 (core)
App: Striim

Hello Erman,

I have mounted an NFS shared drive on CentOS where Striim is running. I have configured several pipelines on Striim (Filereader to Kudu) where these pipelines will read files from the NFS Filesystem.

I notice the VM becomes very slow expecially when reading .gz files.

I have checked the network bandwidth on Zabbix and it is indeed very high.

Capture.PNG

Iftop command also shows the NFS having high read/write.

How can we solve this?

The shared NFS is located on another Datacentre. Should we change the switch?

Thanks,

Roshan

[ErmanArslansOracleBlog]

It may be related with NFS performance actually.. Especially if you started to see that performance degragation after placing that NFS..

You need to declare your requirements first.. the I/O you need from that NFS and then you need to check that NFS and be sure that it can provide that I/O..
Here is a MOS note for checking the NFS perf. Checking NFS Performance (Doc ID 1005928.1)