Start a Conversation

Unsolved

This post is more than 5 years old

F

3074

September 23rd, 2016 03:00

How to stop a recover session?

Hi!

We are getting recover jobs that stall forever.

"

Total estimated disk space needed for recover is 3332 MB.

Requesting 1 file(s), this may take a while...

Recover start time: Wed 21 Sep 2016 06:05:02 AM CEST

53365:recover: NSR server `nebrserver1': busy

71740:recover: Waiting 30 seconds then retrying

.

.

.

"

How do I stop the recover session in a clean fashion?

Using kill or just nsr_shutdown creates side effects .

Namely the device will not be mounted.

Thu 07:34:49 PM Media Info: file disk nebrars1.001 was being written before crash

Thu 07:34:49 PM /nsrstore_0 Cannot seek on /nsrstore_0: Invalid argument

I can only mount the device after running scanner on it. Which takes a long time.

Any ideas?

Thanks

Fernando Silva

9 Posts

September 23rd, 2016 05:00

Hi!

I was working at the server anyway.

The device had not been set to "service mode/disabled"

It could even take backups.

I will try jobkill the next ocasion I get jobs in a strange state.

But anyway if there is a correct way to kill jobs I would expect nsr_shutdown to use it.

But aparently that is not the case, I just had restarted gst and networker - no manual kill - but I still got:

Thu 07:34:49 PM Media Info: file disk nebrars1.001 was being written before crash

Thu 07:34:49 PM /nsrstore_0 Cannot seek on /nsrstore_0: Invalid argument

Thanks

Kind Regards

2.4K Posts

September 23rd, 2016 05:00

Killing the process should already do the job.

In the latest releases you will also find a jobkill command - but this must be executed at the NW server.

However, it is possible that the device has been set to service mode/disabled.

In this case NW will of course not be able to recover.

This is usually caused when you lose control/connection.

Otherwise NW will just wait for the media - please look for an appropriate alert in the admin console.

However, requesting a scan prior to a mount sounds weird - which OS/ NW version do you use btw?

2 Intern

 • 

14.3K Posts

September 23rd, 2016 13:00

To me it looks like your device file is a bit corrupted.

9 Posts

September 26th, 2016 01:00

Hi!

The point is that this occurred in 2 separate although similar systems, after the same sequence of events.

how would you verify if there is corruption?

how would you fix it?

Thanks

Regards

Fernando Silva

25 Posts

September 26th, 2016 04:00

You could also delete the NSR recover task resorce from nsradmin.

If the recover task has gone stale/corrupted, you might find the restore going into infinite loop.

nsradmin

show name

p type:NSR Recover task

Use above 3 commands to verify that you recover job is still there in DB even after aborting it.

If yes, delete it and recheck.

No Events found!

Top