Our Netbackup 188.8.131.52 running on RedHat Linux 6.1 has been crashed two times whithin 1 week, after stable for quite a long time. What should I check? Any experience on this? (I also log the case to Symantec but it is been a little slow.
It is the NBU crashed or the OS itself crashed. If OS is crashing then there is nothing NBU or Symantec can do. If you are saying that it is NBU crashing then please tell us more about it. What is the error message it gives,what is the behaviour, your system logs, is anything recently changed which might have caused this. Normally on windows we check evient viewer if application crashes not sure on Linux but check the system logs which might tell what is crashing.
Given the new EEB support policy I would strongly suggest to reproduce this behavior using 184.108.40.206
Even if support find a bug in code, Symantec will provide EEB for the latest version only
Processes not running? If so, which ones?
As mentioned system messages is a good place to look. Also ensure core logging is enabled at OS level so if any processes are core dumping we can tell which process/es
Thank you for the responses, I am so sorry for the late reply. The crashed was only the Netbackup process the system/server (which is RedHat 6.1) was running fine. Actually I brought up to the Symantec we found 2 coredump files but according to them the coredumps didn't tell anything. So I wait for another crash for the coredump.
The second hung because it is in the cluster the service nbu was autorestart.
Anything to check?
As per my previous post:
Error messages, /var/log/messages, screenshots, processes running / not running will be helpful.
We need to know which NBU processes crashed so that we can tell you which NBU logs are needed.
Cluster restart is a result of NBU processes not running or hanging.
Seems your previous issue is also not solved then?
You need to capture.
The core file itself
A copy of the binary (eg, if bptm crashed, you want the bptm binary)
The system messages log
The log that matches the binary that crashed
Debug output from the core file by following this TN.
This has to be run from your system, any other system will have different library files, and will give the incorrect results.
nbsu -c -t output
It is true, the core file alone is no use, most importantly we need the debug info as explained.
For whatever process it is, please turn the verbose level up, either to 5 if legacy log of Debug /Diag 6 if unified log. If process is nbrb, only turn Debug upto 4.
To set logs :