Comments on: Client inactivity timeout is not group timeout

By: Otmanix' Blog » EMC Networker cumulative hotfix 7.5.2.3 und 7.6.0.5

Otmanix' Blog » EMC Networker cumulative hotfix 7.5.2.3 und 7.6.0.5 — Fri, 04 Jun 2010 12:57:24 +0000

[...] Mehr zum Thema timeouts bei Networker finder sich bei nsrd.info. [...]

By: Otmanix

Otmanix — Fri, 04 Jun 2010 06:32:03 +0000

The hotfix mentioned before is now included in Networker 7.5.2.3:
NW116691: Savegrp did not end despite inactivity timeout reached

ftp://ftp.legato.com/pub/NetWorker/Cumulative_Hotfixes/7.5/cumulative_7.5_readme.txt

By: Otmar Meier

Otmar Meier — Thu, 06 May 2010 09:57:38 +0000

My problem has been solved. I got a hotfix for nsrjobd from EMC.

By: Ondrej Palkovsky

Ondrej Palkovsky — Wed, 28 Apr 2010 15:49:36 +0000

Otmar, this can be fairly easily solved by setting keepalives to something else then the default 2 hours. I think it matters on the storage node actually. It might be enough setting it through NSR_KEEPALIVE_WAIT variable, but we did some testing only with the OS defaults.

As far as client inactivity is concerned, this settings has quite a lot of impact when you use custom savedb scripts. First, the custom save commands implementation by NetWorker is simply flawed; I hope this will be somewhat addressed in 7.5.2.2. However, if you consider the pre-save, then save and then post-save, you cannot heartbeat the pre-save and post-save parts. And you can get quite time-consuming tasks run there without any ability to implement some heartbeat. And that’s why I think we won’t be able to do away with the client inactivity timeout :)

By: Otmar Meier

Otmar Meier — Wed, 28 Apr 2010 07:35:24 +0000

I have an open call similar to this topic.

When a Windows client reboots during backup the networker server doesn’t recognize this situation and keeps the savegroup running and running… You may stop the savegroup but jobsdb still has wrong infos for the rebooted clients. The group will hang again next run. I had to stop nw server, clear /nsr/res/jobsdb/* and start nw server to clean up the situation.

This sometimes happens in our firm when backup overlaps with scheduled task “reboot Windows” (in order to apply updates/patches from WSUS).
This can happen when people from different IT groups do scheduling tasks (backup and reboot) on the same machine.