Unsolved
This post is more than 5 years old
9 Posts
0
730
How to give to a recover session a timeout or some sort of limited attempts.
Hi!
sometimes we get into this situation:
Recovering 1 file from /opt/sdf/backup/ into /opt/assurance
Total estimated disk space needed for recover is 2881 MB.
Requesting 1 file(s), this may take a while...
Recover start time: Mon 04 Jul 2016 01:01:06 PM CEST
53365:recover: NSR server `nebrs1': busy
71740:recover: Waiting 30 seconds then retrying
53365:recover: NSR server `nebrs1': busy
The cause of this situation is not the point of this discussion.
The point is is there a way in the recover command to limit the number of 30 waits of the session or to give it a timeout after which the recover would exit automatically gracefully.
"Gracefully" means not leaving the storage device/volume in an incoherent state.
Thanks
Fernando Silva
ble1
2 Intern
2 Intern
•
14.3K Posts
0
September 23rd, 2016 13:00
There is no retry limit on restore available. If you do ctrl+c you should be fine (you can also kill recover command). Way to achieve what you want is to send log of your session to log file and the parse... it last line = your error code and repeats itself X times; then kill()...
FSSilva1
9 Posts
0
September 26th, 2016 01:00
Hi!
thanks for the reply. Checking the script and issueing the kill would be an approach, but the problem is the kill.
From my experience, the kill will leave the storage device in an sort of "open for recover" state if networker is stoped and started , the device will not be mounted without a scanner.
regards
Fernando Silva
ble1
2 Intern
2 Intern
•
14.3K Posts
0
September 26th, 2016 05:00
If device is not used by any operation, you can reset it via nsrmm -H -f .