constant nfs exports requests

[vkhera]

On my NFS filer (FreeNAS) which is used by proxmox for backups and holding ISOs, I see this in the logs every 10 seconds per PVE host:

Oct 12 17:28:28 filer filer mountd[2262]: export request succeeded from 192.168.7.16
[SIZE=2]

[/SIZE]

Is there some reason for so many export requests from PVE? This goes on all day long. It is mostly harmless, other than cluttering up the logs for finding real problems.

pve-manager/4.0-48/0d8559d0 (running kernel: 4.2.2-1-pve)

 

[manu]

Without going into details, I suspect this is because of the pvestatd: we have to poll regularly that the NFS server is alive, what is it's current space capacitiy so we can report it in the GUI.

 

[vkhera]

I see... looks like you are using this data mostly to populate the RRD graph of available space.

Is it really necessary to poll that frequently given that the RRD graphs are not at such high detailed resolution? It just seems overly chatty on an already busy network.

 

[cpzengel]

this is really embarrassing when you are looking for a serious problem and everything is flooded with stats

 

[RhyvenNZ]

"Can you find out what happened at about 6:20pm tonight?"

Literally two hundred and forty export requests between 6:20 and 6:30. Dang.

 

[Mr.Gosh]

Ran into the same problem today - hard to find the real problems if the logs are spammed

 

[InQuize]

Still no way to disable? These messages are equal to spam...

 

[Christian Jacibsen]

There's a FreeNAS feature request to solve this (https://bugs.freenas.org/issues/15491).
It's not perfect, but it's something.

I'm still digging the Proxmox source code to find the cause.

 

[manu]

Hi
Every 10s seconds, we check the status of the NFS server with 'showmount' and then gather statistics about disk usage *if* 'showmount' successfull.
We need to report early if a NFS server is down, hence the regular check.