What might be happening?

Hello all,

Just having this quite annoying issue in OMV with NFS:

It seems that NFS is having a drastically stop in write copy to disks . Any idea on why this is hapenning?

Or what arguments I could use to solve it?

Kind regards

I have tried all the sysctl recommendations that I found in here: 10Gbe / NFS Tuning

But still having the same issue, no clue, on why this is happening. I though it was related to mergerFS, but no, I have bypassed mergerFS and still having the same issue

BTW, my server has 8VCPUs and 16GB RAM

Any sort of idea?

Hello,

So I think I might know what the issue is. The graph you are seeing on my first message was a network monitor graph. After running that one, I decided to go and start monitoring the disks, as 10gbps which is showing the graph is not the speed a spinning disk writes on itself.

I discovered that when that network speed goes drastically down, the HDD is still copying stuff. What makes me think that NFS caches on RAM most of the big file I am sending through the network and then it passes it to the disk. However if NFS has some sort of memory limit to be used, then it just simply stops, until some free cache space is freed up.

Now the question is: Is there any way to force NFS to directly copy the file to the HDD without caching, so that the network speed is constant and no so spiky.

Kind regards