qmrestore with problems on Proxmox VE 2.0

[cesarpk]

What tar command do you use to test (please use absolute paths)?

Hi dietmar
In the previus reply i showed the commands used, do you have any idea what happens with qmrestore?

Thank you in advance for your attention and your time.

Best Regards
Cesar

 

[dietmar]

In the previus reply i showed the commands used, do you have any idea what happens with qmrestore?

No, sorry. Is it possible to get a copy of such backup file to do more test here?

 

[cesarpk]

No, sorry. Is it possible to get a copy of such backup file to do more test here?

yes, it is possible.

But i get the error message (tar: write error) with several virtual machines:

1- VMs backuped (with vzdump) on the same host on Proxmox VE 2.0
2- VMs backuped (with vzdump) on Proxmox 1.9
(Note: with proxmox VE 1.9 vzdump and qmrestore work fine, my problem is only with qmrestore of Proxmox ve 2.0)

What can we do?, i will wait your instructions.
And thank you very much
Cesar

 

[dietmar]

yes, it is possible.

Where can I download the file (send instructions to <dietmar@proxmox.com>)?

 

[cesarpk]

Where can I download the file (send instructions to <dietmar@proxmox.com>)?

Hi dietmar

qmrestore work fine on Proxmox VE 2.1
Apparently was a bug, with Proxmox 2.1 ended the problem of error message

Best regards
Cesar

 

[derringer]

Hi dietmar

qmrestore work fine on Proxmox VE 2.1
Apparently was a bug, with Proxmox 2.1 ended the problem of error message

Best regards
Cesar

I am getting this same "tar: write error" upon restoring a backup made in proxmox 1.9 to a proxmox 2.0 base (2.1 updated.) .. I only get it when I try to restore one VM, not another. On a simple Win 7 VM with a 100GB hard drive, qmrestore runs without an error. On a different Win 7 VM with one 200GB disk and one 400GB disk, I get this "tar:write error" at the end of qmrestore, but everything seems to work fine.

Is this a fatal error? The VM seems to run fine... any idea if I can safely ignore?

 

[dietmar]

Is this a fatal error? The VM seems to run fine... any idea if I can safely ignore?

Is it possible to get that file (in order to do more tests here)?

 

[moseby]

I see the same problem. I copied the backup files from my box running VE 1.9, did a fresh install of 2.1, and each backup file that I restore gives that error.
They seem to run. No disk failures or space issues.

It is odd that qmrestore does this but vzrestore does not.

For example:

root@hv:~# qmrestore vzdump-qemu-114-2012_06_02-18_17_01.tgz 114
extracting archive '/root/vzdump-qemu-114-2012_06_02-18_17_01.tgz'
extracting 'qemu-server.conf' from archive
extracting 'vm-disk-ide0.raw' from archive
Formatting '/var/lib/vz/images/114/vm-114-disk-1.raw', fmt=raw size=32768
new volume ID is 'local:114/vm-114-disk-1.raw'
restore data to '/var/lib/vz/images/114/vm-114-disk-1.raw' (34359738368 bytes)
tar: write error
34359738368 bytes copied, 68 s, 481.88 MiB/s


Here is a curiocity that may be a clue:

root@hv:~# tar ztvf vzdump-openvz-103-2012_06_01-02_00_01.tgz |head -10
drwxr-xr-x 0/0 0 2010-01-09 11:38 ./opt/
drwxrwxr-x 0/0 0 2011-09-27 22:24 ./opt/bitrock/
drwxrwxr-x 1000/1000 0 2011-04-05 23:53 ./opt/bitrock/zenoss-stack-2.5.1-0/
-rwx------ 1000/114 3296950 2010-01-09 11:44 ./opt/bitrock/zenoss-stack-2.5.1-0/helperBinaryUninstall
drwxr-xr-x 0/0 0 2011-09-27 22:29 ./opt/bitrock/zenoss-stack-3.2.0-0/
-rwx------ 0/0 3906897 2011-08-25 21:45 ./opt/bitrock/zenoss-stack-3.2.0-0/helperBinary
-rwx------ 0/0 1197972 2011-09-27 22:29 ./opt/bitrock/zenoss-stack-3.2.0-0/helperBinaryUninstall
drwxr-xr-x 0/0 0 2011-09-27 22:24 ./opt/bitrock/zenoss-stack-3.1.0-0/
-rwx------ 1000/114 3850601 2011-04-05 23:57 ./opt/bitrock/zenoss-stack-3.1.0-0/helperBinaryUninstall
drwxr-xr-x 0/0 0 2012-06-01 02:00 ./etc/
tar: write error
root@hv:~#

 

[dietmar]

Here is a curiocity that may be a clue:

root@hv:~# tar ztvf vzdump-openvz-103-2012_06_01-02_00_01.tgz |head -10

If you abort the pipe after 10 files (head -10) you get a write error, that is quite obvious. How does that help?

 

[moseby]

I think it helps because if I do the same thing on my desktop (ubuntu 10.04 LTS) it doesn't occur.

moseby@gandalf:~/Desktop/ProxMox$ tar ztvf vzdump-openvz-111-2012_06_01-02_11_20.tgz |head -10
-rw-r--r-- 0/0 0 2012-01-03 23:14 ./.autorelabel
-rw-r--r-- 0/0 0 2012-01-04 00:46 ./halt
drwxr-xr-x 0/0 0 2011-05-11 07:58 ./opt/
drwxr-xr-x 0/0 0 2012-06-01 02:11 ./etc/
drwxr-xr-x 0/0 0 2012-01-03 23:31 ./etc/lvm/
drwx------ 0/0 0 2011-09-02 12:21 ./etc/lvm/backup/
-rw-r--r-- 0/0 24874 2011-09-02 12:20 ./etc/lvm/lvm.conf
drwx------ 0/0 0 2011-09-02 12:21 ./etc/lvm/cache/
drwx------ 0/0 0 2011-09-02 12:21 ./etc/lvm/archive/
-rw-r--r-- 0/0 112 2007-01-07 05:44 ./etc/mail.rc
moseby@gandalf:~/Desktop/ProxMox$

 

[moseby]

Further, I found this http://archives.linuxfromscratch.org/mail-archives/lfs-support/2010-March/038236.html
The mailing indicates that the write error is caused by a change in tar. They quote the same two versions that I used. As I said before, it does throw the error message but it doesn't seem to have caused any harm.

ProxMox
------------------
root@hv:~# tar --version
tar (GNU tar) 1.23
Copyright (C) 2010 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>.
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.

Written by John Gilmore and Jay Fenlason.


Desktop
------------------
moseby@gandalf:~/Desktop/ProxMox$ tar --version
tar (GNU tar) 1.22
Copyright (C) 2009 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>.
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.

Written by John Gilmore and Jay Fenlason.

 

[dietmar]

Further, I found this http://archives.linuxfromscratch.org/mail-archives/lfs-support/2010-March/038236.html
The mailing indicates that the write error is caused by a change in tar. They quote the same two versions that I used.

Is that bug already reported on the Debian BTS? If not, please can you report one?

 

[moseby]

I poked around in the Debian git server and saw that updates had been made to tar concerning signal usage/handling. I downloaded the latest tar version, 1.26, and confirmed that it does not throw the write error. So it appears to be fixed, we just need a version update.

 

[dietmar]

I downloaded the latest tar version, 1.26, and confirmed that it does not throw the write error. So it appears to be fixed, we just need a version update.

Will they move that version into the stable debian repository?

 

[Jérôme]

hi

Let me know, everyone has this bugs used ".qcow2" disks ?

I use ".qcow2" disks on KVM 2K3 Server and it happening.

So, I ask you this cause with qcow it's provisionning and "tar" or "qmrestore" can be crash about this.

...

I hope it can help

Best regards.