Cluster node acting funny

[masterdam79]

Hi all,


Got a two-node cluster setup from which one of the cluster nodes is acting funny.

Tried to google the problem of course but it seems ingoogleable ;-)


Here is the situation:

Cluster node "proxmox1": 10.0.2.50 - problem case - pve 3.1
Cluster node "proxmox2": 10.0.2.26 - running fine - pve 3.1

On this node the /etc/pve mountpoint is only mounted momentarily for a very brief moment, just enough to start or stop a vz container with "pmxcfs && vzctl start|stop <ID>".

Therefore when I log in to the web-interface using https://10.0.2.26:8006 I can see all VMID's from both cluster nodes, but all info for the VM' s from proxmox1 is not showed, just the VMID's/VZID's.

The VZ's/VM's are running fine when I start them manually using above command combined with "pmxcfs".


I tried the following:

1) Reboot cluster node
2) Reboot both cluster nodes simultaneously
3) restart cman
4) restart pve-cluster
5) restart pvestatd
6) restart pvedaemon
7) restart pve-manager


Logfile:
/var/log/syslog is full of entries like:
Feb 15 19:27:20 proxmox1 pvestatd[2958]: WARNING: ipcc_send_rec failed: Connection refused


Anyone got any ideas?

If more info is needed from any logfile let me know.

I'm inclined to reinstall the full cluster node and restore my backups once I have re-entered the node into the cluster.

Thanks in advance!



Richard


Verstuurd vanaf mijn ASUS Transformer Pad TF300T met Tapatalk

 

[udo]

Hi,
sounds like you have trouble with multicast on the switch.

Can you enable multicast?

Otherway switch to unicast - see here to find the issue:
https://pve.proxmox.com/wiki/Multicast_notes

Udo

 

[masterdam79]

Hi Udo,

Thanks, I'll look into that in my router (Draytek Vigor 2920n), got only dumb switches in the house.

Verstuurd vanaf mijn ASUS Transformer Pad TF300T met Tapatalk

 

[masterdam79]

I did the multicast ping test but all seems well, sending and receiving packages back and forth..

Verstuurd vanaf mijn ASUS Transformer Pad TF300T met Tapatalk

 

[masterdam79]

omping & ssmping from proxmox1 to proxmox2 and visa versa

Verstuurd vanaf mijn ASUS Transformer Pad TF300T met Tapatalk

 

[masterdam79]

Started reinstalling full cluster btw

Verstuurd vanaf mijn Nexus 5 met Tapatalk

 

[mir]

Your /etc/hosts file on both nodes has declaration of both nodes?

10.0.2.50 proxmox1 proxmox1.foo.bar
10.0.2.26 proxmox2 proxmox2.foo.bar

 

[masterdam79]

Hai Mir,

It has been working fine for a couple of months and suddenly overnight it wasn't anymore.

If my hosts files weren't correct, which I assume they were, I had also configured them in DNS and RDNS.

The only two things I can think of that might have had something to do with it suddenly not working anymore are:

1 - Backup locations change from NFS to local (/var/lib/vz/dump)
2 - Nagios NRPE configuration

Other than that, nothing changed in my setup.

Thanks

Verstuurd vanaf mijn Nexus 5 met Tapatalk