Snmp timeout when nfs storage is down.

[max.nolent]

Hello,

I have the same bug on few proxmox cluster, i have NFS storage to do some saves of containers. Snmp works fine but every time my nfs storage goes down, snmpd had timeout. If i force to unmount the nfs storage, the snmp work again. Have you any idea to configure snmpd to stop crash when nfs storage gone ?

 

[bbgeek17]

SNMP as in Simple Network Management Protocol ?

https://en.wikipedia.org/wiki/Simple_Network_Management_Protocol

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[max.nolent]

Yes snmp as simple network management protocol

 

[bbgeek17]

Given that SNMP is not part of standard Proxmox installation, your question does not fit the topic of the forum "Proxmox VE: Installation and configuration". Both of the involved systems (NFS and SNMP) are not provided by PVE/Proxmox. You may get general help here in the forum but you are going to provide much more information than "times out".

A random guess - your SNMP, wherever its running, is probing NFS and timing out because it cant?

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[max.nolent]

I have other server(Not Proxmox but debian 9,10,11) with NFS and SNMP on it, i don't have any problem, if my storage goes down. I get timeout or critical only on storage involve. With proxmox OS, when my nfs goes down, all request of snmp timeout (memory,cpu,network card etc...) i can't even contact the daemon SNMP. Its a particularity of proxmox, so i think my question is on the good forum, if any person had already the problem and know how to resolve it, I will thank you for it. I must specify that i have the problem on all my proxmox cluster.

 

[bbgeek17]

Ok, now that you provided much more insight into how the setup is done and where pieces are located, my next guess is that its the artifact of how Proxmox is mounting the NFS: ,hard,proto=tcp,timeo=600,retrans=2,

I suspect SNMP IO probing is simply hung waiting for response, as it should be on such NFS mount. Are your other systems using hard mounts?
If the IO wait is the issue for SNMP you will need to figure out how to make SNMP give up at application layer faster.
You can google for "snmp hard mount nfs"

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[max.nolent]

Thanks you for your answer, i put the option "soft" in /etc/pve/storage.cfg, that would resolve my problem.

 

[bbgeek17]

I am not sure if you meant it as a statement or a question. I would strongly advise against using soft mount if you are storing VM disks on your NFS. In that case you will likely have data corruption within VM disk at some point.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[max.nolent]

I only use NFS to saves my container, i don't have VM disk on this storage.

 

[bbgeek17]

I only use NFS to saves my container, i don't have VM disk on this storage.

the only data that is somewhat safe is read-only one, ie ISO.
everything else is bound for corruption.

its your choice, do as you wish

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox