If you have a core file in /usr/OV/PD/cores/netmon then you can be sure it
is a netmon core file, because we had to add code to make sure it ended up
there. Otherwise it would be in root (/).
There are several ways you can look at a core you don't know the origin of,
like strings or od, to try to determine its origin, but they are of little
use to you in determining what actually happened. On AIX V4 you can often
enter
od -c <core_file_name> 0x6b0 | head
and get it to print out a hex array which spells out the failing process.
With strings you have to guess. Usually it is the first executable file
name you find in the output.
Then you need to use a core analysis tool which is matched to the compiler
that was used to produce the executable. On AIX, for NetView, that is dbx.
On Solaris it is gdb. On Digital it is DEC's ladebug. They all have about
the same syntax,
something like <tool_name> <core_file_name> <executable_name>. For
netmon on AIX that would be
dbx /usr/OV/PD/cores/netmon/core /usr/OV/bin/netmon
and when the prompt comes back you can issue commands like "where" to give
you a stack trace and "registers" to show you the registers and so on.
But unless you have access to the actual source code, it won't mean much.
Finally, for AIX, we ship a core analysis tool /usr/OV/service/readcore
which uses dbx (dbx must be installed on your AIX system to use readcore).
If you call Support with a core, they will ask you to run this. We have
a similar tool for Solaris called readcoreg, but it is not shipped with
the product. You must download it and gbd from our ftp site to run it. We
have nothing pre-packaged for Digital, since their tool, ladebug, doesn't
show you as much as the others if you don't have a fully debuggable module.
So if you have a netmon core on AIX, cd to /usr/OV/PD/cores/netmon and run
/usr/OV/service/readcore core /usr/OV/bin/netmon
or just /usr/OV/service/readcore and let it take the defaults. It
should produce a core.report in that directory which will give you an
analysis of that core that you can use when you contact Support.
James Shanks
Tivoli (NetView for UNIX) L3 Support
Phat D Doan <pddoan@REGENCE.COM> on 03/19/99 11:47:31 AM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on
NetView <NV-L@UCSBVM.UCSB.EDU>
To: NV-L@UCSBVM.UCSB.EDU
cc: (bcc: James Shanks)
Subject: Re: Netmon always dying
How I can view/read the core file i.e. /usr/OV/PD/cores/netmon/core?
Thanks.
Phat Doan
From: James Shanks <James_Shanks@TIVOLI.COM> on 03/19/99 06:04 AM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on
NetView <NV-L@UCSBVM.ucsb.edu>
To: NV-L@UCSBVM.ucsb.edu
cc: (bcc: Phat D Doan/BCBSO/TBG)
Subject: Re: Netmon always dying
A signal 11 is a core. You must have a core file in
/usr/OV/PD/cores/netmon. Look for it. If you find one, call support back
and report that. There were several netmon cores fixed in 5.1.1. There
was avery early fix package for one of them.
James Shanks
Tivoli (NetView for UNIX) L3 Support
Jean-Michel THIEBAUD <jean-michel.thiebaud@DETEXIS.THOMSON-CSF.COM> on
03/19/99 02:59:44 AM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on
NetView <NV-L@UCSBVM.UCSB.EDU>
To: NV-L@UCSBVM.UCSB.EDU
cc: (bcc: James Shanks)
Subject: Netmon always dying
Hi,
I work on Netview V5.1 for UNIX . My oslevel system is AIX 4.2.1. I
have 512 Mo RAM and 1Go SWAP on a H50 CPU. I don't know why but the
daemon netmon is always dying. it died always exiting on level 11. I
have seen in this forum someone saying that's there was some patches to
correct this problem.. is it true ? I called my Tivoli support yesterday
but they told me there are not.
Any idea about this problem ?
Thanks
Jean-Michel THIEBAUD
THOMSON-CSF DETEXIS
|