Replies to This Discussion

Permalink Reply by Jerry Schmidt on June 12, 2014 at 6:30am

Hi SistemistiBucap,

I checked the screenshot and the backup is failing on a pancbt.vmdk file. Have you tried disabling the CBT option and just backing up with the EXPORT option enabled?

• ▶ Reply

Permalink Reply by SistemistiBucap on June 12, 2014 at 6:52am

Hello Jerry,

I've tried to follow your suggestion disabling the CBT option and leaving the vm exported but it failed again with the same 'Return code: 5'. It failed after about 3 minutes. We are backing up 5 vm (4 win+1 linux) and only the linux one give us this kind of error. Any other trick is welcome!!

Thank u for your check

Fabio

Attachments:

• screen_shot_2#.png, 123 KB

• ▶ Reply

Permalink Reply by Jerry Schmidt on June 12, 2014 at 7:00am

Trying selecting the VM and disable the EXPORT option, then power cycle the vmPro node. After vmPro is back online, select the VM again and enable the EXPORT option and try the backup again.

If the backup fails again go to Help-vmPRo System and look at the end of the controller log and vm_proxy_fs log. Copy that output and post it here, this should give us a better idea of why the backup continues to fail.

• ▶ Reply

Permalink Reply by SistemistiBucap on June 12, 2014 at 7:29am

Yes, unfortunately it fails again ...

In the attachments you could find that log, thanks  

Attachments:

• vm_proxy_fs.log, 15 KB

• ▶ Reply

Permalink Reply by Jerry Schmidt on June 12, 2014 at 7:43am

We will look at it. If possible, attach the controller log too. 

• ▶ Reply

Permalink Reply by SistemistiBucap on June 12, 2014 at 7:49am

Log attached

Attachments:

• controller.log, 11 KB

• ▶ Reply

Permalink Reply by Jon McNerny on June 12, 2014 at 7:58am

Appears to be a problem with PVS handling the partition.  I checked and PVS is linked to LVM.

bash-4.1# ls -alh /sbin/pvs

lrwxrwxrwx 1 root root 3 Jan 15 2013 /sbin/pvs -> lvm
bash-4.1#

Logging shows a problem with our filesystem API not handling the partition well.

2014-06-12 16:18:23.091350: ERROR [thread:140249137149696, where:spawn.c:246, error:5] /sbin/pvs
vm_proxy_fs[2757]: ERROR [thread:140249137149696, where:spawn.c:246, error:5] /sbin/pvs
2014-06-12 16:18:23.154208: ERROR [thread:140249137149696, where:vm_proxy_api.c:436, error:2] /test/Linux_10.0_Jdweb/Linux_10.0_Jdweb-flat.vmdk: No such file or directory
vm_proxy_fs[2757]: ERROR [thread:140249137149696, where:vm_proxy_api.c:436, error:2] /test/Linux_10.0_Jdweb/Linux_10.0_Jdweb-flat.vmdk: No such file or directory

Please SSH into your linux box and give us some disk partition details.  Possible the stdout of 'fdisk -l', 'parted -l' could be helpful.

Thanks,

Jon

• ▶ Reply

Permalink Reply by SistemistiBucap on June 12, 2014 at 8:25am

In my linux box there isn't any /sbin/pvs file

It's an old Debian 3.0 long time not upgraded 

I could give you the spool of the 'fdsik -l' command (screenshot attached), 'parted'  is not present as command

Attachments:

• fdisk_sshot.png, 13 KB

• ▶ Reply

Permalink Reply by Jon McNerny on June 12, 2014 at 8:34am

The /sbin/pvs is our binary to work with the virtual machines file systems/partitions. 

Can you manually snapshot this in VMware?

I'm fairly certain we don't support 'Raid auto detect' or maybe the problem is this 'Win95' filesystem causing our API to error out.

• ▶ Reply

Permalink Reply by SistemistiBucap on June 12, 2014 at 8:43am

Yes, if I try to make a snapshot inside Vmware (by vcenter) it works fine.

As you could see in the vm_proxy_fs.log, the 'snaphsot' step, during the vmpro backup, works fine too:

2014-06-12 16:18:06.67989: Creation of snapshot for VM 'Linux_10.0_Jdweb' completed.

• ▶ Reply

Permalink Reply by Jon McNerny on June 12, 2014 at 8:46am

Probably our API then.  If you have a support contract I'll be happy to open a ticket and pursue the bug further.

Feel free to email me if so. jonathan.mcnerny@quantum.com.

Thanks for bringing it our attention.

Jon