We have recently upgraded our primary backup servers from Networker version 7.2.2 to version 7.4.3 – we are having a number of problems with the new version:
1] Groups frequently stall at ~98% – 99% complete and need to be stopped and rerun to complete. This is an issue since we cannot leave Networker to do unaided backups any more because the groups would remain in a stalled state indefinitely and subsequent backups are then skipped.
I have tried clearing the /nsr/tmp directory and restarting the service as well as reducing the jobsdb settings to 4 days retention with a database size of 200,460 KB, however, the problem persists.
P.S. Stopping a group under new GUI always takes 2 attempts before accepting the command – this may be a feature but no prompt is returned to this effect.
2] Group rerun issues
i) There is a difference in the way group reruns work in the new version – it appears that a pair of parameters, Interval and Restart window affect this.
By default the Interval is 24:00 hours and Restart window 12:00 hours – thus, if we restart the group over 12 hours after it started it is rerun rather than restarted.
I have increased the Restart window to 23:59 to allow a rerun up to 24 hours after starting but our full Friday night backup sometimes exceeds 24 hours and as a result cannot be restarted to capture failed clients – this worked fine on 7.2.2. I’m not sure what function the Interval parameter serves since the client schedule determines the frequency of backups.
ii) Our daily schedule is for an incremental backup daily and a full backup on Friday. If failures are detected on the Thursday night incremental backup and it is rerun on Friday morning, Networker performs a full Friday backup of the failed savesets rather than the incremental Thursday backup we were attempting to rerun.
iii) On a rerun all the client indexes appear resaved regardless of whether they previously succeeded – on the old release only the indexes of the unsuccessful clients were saved again on a rerun.
3] Relabelling expired volumes
We use the nsrjb –j <jukebox> –L –R <volume> command in batch scripts to move older tapes from our daily pools into monthend pools. The –R option is supposed to prevent the user from being prompted if the volume being labelled is recyclable or expired. This no longer appears to work and the user is prompted regardless. This was a useful feature to prevent accidental overwriting of valid tapes, if the operator was prompted they answered ‘no’ as Networker had determined that the selected volume was still in use.
Thanks!
1] Groups frequently stall at ~98% – 99% complete and need to be stopped and rerun to complete. This is an issue since we cannot leave Networker to do unaided backups any more because the groups would remain in a stalled state indefinitely and subsequent backups are then skipped.
I have tried clearing the /nsr/tmp directory and restarting the service as well as reducing the jobsdb settings to 4 days retention with a database size of 200,460 KB, however, the problem persists.
P.S. Stopping a group under new GUI always takes 2 attempts before accepting the command – this may be a feature but no prompt is returned to this effect.
2] Group rerun issues
i) There is a difference in the way group reruns work in the new version – it appears that a pair of parameters, Interval and Restart window affect this.
By default the Interval is 24:00 hours and Restart window 12:00 hours – thus, if we restart the group over 12 hours after it started it is rerun rather than restarted.
I have increased the Restart window to 23:59 to allow a rerun up to 24 hours after starting but our full Friday night backup sometimes exceeds 24 hours and as a result cannot be restarted to capture failed clients – this worked fine on 7.2.2. I’m not sure what function the Interval parameter serves since the client schedule determines the frequency of backups.
ii) Our daily schedule is for an incremental backup daily and a full backup on Friday. If failures are detected on the Thursday night incremental backup and it is rerun on Friday morning, Networker performs a full Friday backup of the failed savesets rather than the incremental Thursday backup we were attempting to rerun.
iii) On a rerun all the client indexes appear resaved regardless of whether they previously succeeded – on the old release only the indexes of the unsuccessful clients were saved again on a rerun.
3] Relabelling expired volumes
We use the nsrjb –j <jukebox> –L –R <volume> command in batch scripts to move older tapes from our daily pools into monthend pools. The –R option is supposed to prevent the user from being prompted if the volume being labelled is recyclable or expired. This no longer appears to work and the user is prompted regardless. This was a useful feature to prevent accidental overwriting of valid tapes, if the operator was prompted they answered ‘no’ as Networker had determined that the selected volume was still in use.
Thanks!