Snapvault: relationship is in wrong status: Idle

[dailyfresh]

Hello,

Local backups are fine. But on al ten datasets we get the following error in the DPM Job log:

'filer02:/Volume_backup: Could not get completion time for the data transfer from filer01:/vol_volumename/qt_name to filer02:/Volume_backup/qt_qtname: relationship is in wrong status: Idle'

Job fails after this error. I can't find any joblog which states the error message in detail.

The 'Notification' section within DPM only states: "A backup of relationship between filer01 and filer02 failed. Please refer to job 1829 for more details." The referred job details are described above.

More info:

Filer01 is located at branch A and filer02 is located at branch B. Connection between the two branches is by a persistent VPN tunnel. All protocols and ports are allowed. Ping latency is below 5ms.

Manual starting snapvault on the second filer "Snapvault update filer02:/Volume_backup/qt_qtname" does work without any errors.

Has someone any experience with the first DPM error message?

[dailyfresh]

I changed to MTU size of the VPN tunnel. Hoped that the wrong status message would be resolved by that. But unfortunalty it doesn't. Running out of ideas...

[dailyfresh]

Looks like a discussion with me, myself oh and I...

Anyway, changed the MTU size on the filer as well. Adjusted ‘ifconfig vif01 `hostname`-vif01 netmask 255.255.255.0’ in ‘ifconfig vif01 `hostname`-vif01 netmask 255.255.255.0 mtusize 1388 trusted wins up’. Link down and up again.

Backup result is the same: Failed. During backup I opened DPM and the console. During transfer both are the same. After the last bytes have been sent, snapvault on console is idle. DPM however isn't. DPM seems to be waiting for something and raises an (the) error after exactly five minutes.

Still hoping for someone with an idea...

[dailyfresh · Accepted solution]

This was a bug (495927) in the DPM software. Bug solved in update version 4.0.2D4.