unreachable nfs causes high load

[limone]

Hi,

I have my local router connected to a proxmox installation via nfs (over vpn).
It's working, I noticed, when my router is down, and so the nfs storage isn't reachable on the proxmox server, the server load goes up from 0.05 (test server, no vm's running) to 1, and from 0.5 to 3.5 (low used server)

Is this a known issue, or someway I can fix this?

 

[wolfgang]

Hi,

I guess this is the reconnects from the NFS, what makes this load.
But to be sure you have to look which process makes this load.

 

[limone]

I'm not entirely sure how to check this, I used 'ps -ax' and the only process with a 'D' is snmp.
I guess it's trying to read to information from that mounted drive, and that doesn't work and the load goes up.

On another server I had a script running every minute to remount the nfs storages, and I had no timeout built in, so the load went up to 200+ and then the server crashed (no ram left)
SNMP seems to have a timeout, but I'm not entirely sure about that, because the load keeps being constat.

Spoiler: nfs-remount.sh

#!/bin/bash
list=$(ls /mnt/pve)

for i in $list
do
status=$(timeout 10 ls /mnt/pve/$i 2>&1)

if [[ $status =~ .*Stale.* ]]
then
umount /mnt/pve/$i
fi
done

So, not a proxmox issue..