Hello
I know, it's an very comun problem and there is a lot of post about this, but I haven't found an definitive solution to the problem.
I have an cluster with proxmox 4.4 of 15 servers, all of them have an NFS share connection to an Freenas server for backup files.
Sometime, for some reason, an server loose connection to freenas (other still have, so it's don't appear be an freenas server, but yes an probably freeze connection of thie proxmox server), and automatically the cluster go down for this server (red light) and KVM hosted in this server begin to have problem.
I am unable to stop the dump in this case...
- From GUI it's not possible, action of stop have no effect
- From ssh, vdump -stop don't solve problem and I can't kill one of the tree vdump process (I kill it, but process still active). I am also not able to access to NFS share directory in this case.from this server (from other server, access still be OK)
The only solution that I found is... to reboot server for now.
I have seen the "umount -t -l " in other item, that I am pending to test the next case, but is there not some other way? I have the NFS connection in soft, not hard, in storage, but in anyway, an NFS freeze locks all server :
mount result :
192.168.1.1:/mnt/backup/VPS on /mnt/pve/FREENAS type nfs (rw,noatime,vers=3,rsize=16384,wsize=16384,namlen=255,soft,nolock,noacl,proto=udp,timeo=60,retrans=5,sec=sys,mountaddr=192.168.1.1,mountvers=3,mountport=801,mountproto=udp,local_lock=all,addr=192.168.1.1)
storage :
nfs: FREENAS
export /mnt/backup/VPS
server 192.168.1.1
path /mnt/pve/FREENAS
maxfiles 5
content images,iso,backup,vztmpl,rootdir
options rw,noatime,vers=3,noacl,rsize=16384,wsize=16384,namlen=255,proto=udp,intr,noatime,timeo=60,retrans=5,async,nolock,noacl,soft
I am pending to review the real problem of freeze (this is the base of problem) but,is there some better configuration option?
Thanks a lot for your help
I know, it's an very comun problem and there is a lot of post about this, but I haven't found an definitive solution to the problem.
I have an cluster with proxmox 4.4 of 15 servers, all of them have an NFS share connection to an Freenas server for backup files.
Sometime, for some reason, an server loose connection to freenas (other still have, so it's don't appear be an freenas server, but yes an probably freeze connection of thie proxmox server), and automatically the cluster go down for this server (red light) and KVM hosted in this server begin to have problem.
I am unable to stop the dump in this case...
- From GUI it's not possible, action of stop have no effect
- From ssh, vdump -stop don't solve problem and I can't kill one of the tree vdump process (I kill it, but process still active). I am also not able to access to NFS share directory in this case.from this server (from other server, access still be OK)
The only solution that I found is... to reboot server for now.
I have seen the "umount -t -l " in other item, that I am pending to test the next case, but is there not some other way? I have the NFS connection in soft, not hard, in storage, but in anyway, an NFS freeze locks all server :
mount result :
192.168.1.1:/mnt/backup/VPS on /mnt/pve/FREENAS type nfs (rw,noatime,vers=3,rsize=16384,wsize=16384,namlen=255,soft,nolock,noacl,proto=udp,timeo=60,retrans=5,sec=sys,mountaddr=192.168.1.1,mountvers=3,mountport=801,mountproto=udp,local_lock=all,addr=192.168.1.1)
storage :
nfs: FREENAS
export /mnt/backup/VPS
server 192.168.1.1
path /mnt/pve/FREENAS
maxfiles 5
content images,iso,backup,vztmpl,rootdir
options rw,noatime,vers=3,noacl,rsize=16384,wsize=16384,namlen=255,proto=udp,intr,noatime,timeo=60,retrans=5,async,nolock,noacl,soft
I am pending to review the real problem of freeze (this is the base of problem) but,is there some better configuration option?
Thanks a lot for your help