Solaris 10 core dump caused by nbsl daemon

[Skywalker1957]

I have a Solaris 10 master, patch rev Generic_127127-11, that is core dumping caused by the nbsl daemon. I'm using Netbackup 6.5.4 and I've got a case into Symantec for it, but haven't gotten a callback yet. Has anyone run into this before? I know there are issues with this with NBU 6.0 and 6.5.2 but this is something new for 6.5.4. I'm trying to get a window to patch him up to rev, but it's not likely anytime soon. We've made no upgrades or substantial changes to the software or environment. The only thing different from last month to now, is maybe about a 20% change in size of my catalog to a little over a TB now. Any ideas or insight is always appreciated. Thanks!

 

[OzzMosiz]

Can't tell you why it's happening. Does it happen often or infrequently?

Can you raise the logging level of your nbsl log to 6 and when it core dumps, note the time of the core and gather the raw log file for nbsl for that timeframe.
Also collect the core and also run a pstack <core> and pfiles <core> and pmap <core> and redirect the output to a file pstack.nbsl.txt for example. Then send that into Symantec - that will give them a headstart in helping I guess!
Do you have NOM running on your Master also?
I understand this can cause issues at times.

 

[Skywalker1957]

According to Symantec this is a problem when NOM queries the nbsl daemon on the master too much. They are going to get me a patch that will correct this and will be released in 6.5.5.