Unsolved
This post is more than 5 years old
1 Message
0
925
NMC failure on Linux Networker 7.5.3.2
Hello,
We recently installed a very large NDMP environment on a RHEL 5.5 Linux server with 48GB RAM, 8 procs (GL 585). What we have been experiencing is our load averages get very high (above 10) and the NMC GUI no longer responds. Is there a known workaround to this like increase QoS UDP priority? As a note, nwadmin, nsradmin do not work when the loads get this high; memory is tapped with all 48GB used and some swapping occurring. We are backing up 6 data movers with 4 streams apiece into 24 drives. 23 drives are in use currently.
Any help would be appreciated.
Thanks,
Colin
MikeToo1
38 Posts
0
September 17th, 2010 07:00
Hello Colin,
Please confirm which nsr service if any is using that much resources.
MikeToo1
38 Posts
0
September 17th, 2010 18:00
Please verify the setting for jobsdb. Depending on the size of the back, the jobsdb default size is 40Mb. If the current size of the jobsdb is at or exceeds the maximum size you will need to review the performance tuning guide. The current confirgured size is under the properties on the NMC of the NW server. Under the second tab your will see the setting for days retained and maximum size.
Compare that size to the size of the jobsdb under \nsr\res\jobsdb. Please raise maximum size to double the current actual size until you can determine what is needed.
ble1
14.3K Posts
0
October 21st, 2010 04:00
Your issue might be related to nsrjobdb. In 7.5.3.3 there is bunch of fixes for nsrjobdb so chances are they might address this issue.
I myself will go to this release this weekend. I used to have 7.4.5.x (7.4.5.8 was the last build I used) and has issues where nsrjobdb would simply go to 100% or more (
) after week or two. I used 400MB size and retention 3 days. I thought this could be related to threading which got managed better in 7.5.x so I didn't bothered myself opening an incident ticket. 2 or 3 weeks ago I upgraded to 7.5.2.4 (as we started client side tests with that release in August) and found that nsrjobdb goes crazy now even before - after 3 days of running it would just sit on 100%. Debugging showed that it fails during purging so I increased retention to 9 days hoping restart and cleanup every weekend would help, but that didn't happen (meaning it is not just purging). I will give 7.3.3.3 a chance and if it doesn't work I will simply make a noise until I get it fixed. I know that in my case we have rather heavy environment which generates 160k entries for nsrjobdb on daily basis, but this is enteprise product so I see no issue with it (most of this entries are caused by database and archive log backups of course).