This post is more than 5 years old
4 Posts
0
891
November 25th, 2014 10:00
Savestream redirect on device fail
Salut, everyone! My Networker datazone is simple but needs to be available In case of device fail. Imagine one client with one saveset, NWserver with 2 locally attached tape libraries, one group and one exact pool to catch all the savesets. Group starts, NW finds mounted tape on 1st library and directs the savestream to nsrmmd daemon servicing the drive. During saveset writing the drive/tape fails, so the group fails too and i don't get any backup at all. But there's a second exact library ready to be used as a destination device to reach another tape from the same pool.
How can i configure networker to automatically redirect saveset on device fail to the working one?
0 events found
ble1
4 Operator
•
14.4K Posts
0
November 25th, 2014 14:00
It deals with any failure - and if failure is caused by tape drive gone broken and this is seen as failure - it will either fail or retry based on this setting. As it fails in your scenario, then retry should do it again and if drive has been marked bad it will use next one available.
AskKuzma
4 Posts
0
November 25th, 2014 12:00
True, next backup won't use the failed device, but my goal is to restart the backup of the failed saveset on the working one, and that doesn't happen automatically for the failed saveset(
ble1
4 Operator
•
14.4K Posts
1
November 25th, 2014 12:00
If drive fails, it will be marked in service mode or disabled and then automatically next drive which matches pool selection criteria will be used.
AskKuzma
4 Posts
0
November 25th, 2014 13:00
For example, Symantec Netbackup has the ability to deal with such a situation using Storage Unit Group functionality: there's no need for manual restart in case of device failure, it simply gets my saveset done... As i mentioned earlier, my system needs to be highly available, so there's no time to wait for backup operator reaction.
ble1
4 Operator
•
14.4K Posts
1
November 25th, 2014 13:00
As restart is manual action, you can add extra step of marking drive as unusable and it will use next drive.
AskKuzma
4 Posts
0
November 25th, 2014 13:00
That's interesting, you mean the "client retries" attribute for a group? the number of times failed clients should be retried before savegrp gives up and declare them failed - I thought it deals only with network problems and unresponsive clients, but if it will catch the "failed device" scenario, i let you know. Many thanks, Hrvoje))))))
ble1
4 Operator
•
14.4K Posts
1
November 25th, 2014 13:00
If you need HA setup, then your devices are spread across multiple media servers and NW supports stream distribution across them. You can also increase group retries to have this automatically handles (given that devices would really be marked as broken in the meantime).