vzdump sometimes fails with exit code 2 when creating backup of KVM machines

H

holgerb

Member
Aug 3, 2009
45
0
6
Hi all,

we have a strange problem when making backups of our KVM machines. We have a cluster consisting of mixed cluster nodes and a central SAN for storing the images of the machines via NFS.

When launching the backup with vzdump we get error messages like this:
Code: 
proxmox-epr005:~# vzdump --node 6 --snapshot --compress --storage BACKUP --mailt                                              o XXX --mailto XXX 116
INFO: starting new backup job: vzdump --node 6 --snapshot --compress --storage B                                              ACKUP --mailto XXX --mailto XXX 116
INFO: Starting Backup of VM 116 (qemu)
INFO: running
INFO: status = running
INFO: backup mode: snapshot
INFO: bandwidth limit: 10240 KB/s
INFO:   Logical volume "vzsnap-proxmox-epr005-0" created
INFO: creating archive '/mnt/pve/BACKUP/vzdump-qemu-116-2010_07_27-10_54_38.tgz'
INFO: adding '/mnt/pve/BACKUP/vzdump-qemu-116-2010_07_27-10_54_38.tmp/qemu-serve                                              r.conf' to archive ('qemu-server.conf')
INFO: adding '/mnt/vzsnap0/images/116/vm-116-disk.qcow2' to archive ('vm-disk-id                                              e0.qcow2')
INFO: 196296704 B 187.2 MB 361.9 s (6:01 min) 542477 B/s 529.76 KB/s
INFO: write: Input/output error
INFO: received signal - terminate process
INFO:   Logical volume "vzsnap-proxmox-epr005-0" successfully removed
ERROR: Backup of VM 116 failed - command '/usr/lib/qemu-server/vmtar '/mnt/pve/B                                              ACKUP/vzdump-qemu-116-2010_07_27-10_54_38.tmp/qemu-server.conf' 'qemu-server.con                                              f' '/mnt/vzsnap0/images/116/vm-116-disk.qcow2' 'vm-disk-ide0.qcow2' '/mnt/vzsnap                                              0/images/116/vm-116-disk-1.qcow2' 'vm-disk-ide1.qcow2' |gzip |cstream -t 1048576                                              0 >/mnt/pve/BACKUP/vzdump-qemu-116-2010_07_27-10_54_38.dat' failed with exit code 2
INFO: Backup job finished with errors
Sometimes restarting the backup for the same machine runs without issues. We are experiencing this behaviour for different machines on different hosts.

Questions:
- Is someone else experiencing similar issues ?
- What could be the reason for this ?
- Is there a list of exit codes for vzdump ?

TIA,
Holger
 
Last edited:
holgerb said:
Yes, more than enough. We are talking about a VM with something like 30 GB HDDs and someting like 3 TB available on the NFS share which is used for backup.
Click to expand...

And that NFS share is writable for 'root'? Any hint in syslog?
 
The NFS share is fully writable for root. The really strangest thing is that a lot of other VM backups to the NFS share work fine and then we are experiencing the problem described.

When I start the same backup job manually the next morning it will most likely run without any problem. The problem is not limited to a special machine. The only pattern I can see is that it occurs more often for VMs with more than one virtual HDD attached. Unfortunately syslog doesn´t give any hint.
 
Since our Thecus is a closed-source platform there is no access to any log files. The protocols available on the web frontend do not reveal anything.
 
You must log in or register to reply here.
Top Bottom