red hat 9 hp agents

[KristofH]

hi,

just wondering when the rh9 agents will become available, because installing the rh8 agents (in particular the ucd-snmp package) doesn't seem to work.

maybe there's a tweak to get them going? just wondering..

cheers..

[KristofH]

anyone else tried installing the rh8 agents on rh9?

[Ramkumar Devanathan]

Maybe some details of error messages shown would help... should typically work most of it...

- ramd.

[Mogens Kjaer]

I've tried installing

hpasm-6.40.0-16.Redhat8_0.i386.rpm
cpqacuxe-6.40-11.0.i386.rpm
cmanic-6.40.0-3.i686.rpm
cmastor-6.40.0-8.i386.rpm
cpq_cciss_RedHat80-2.4.46-7.i386.rpm
cpq_cpqarray_RedHat80-2.4.26-7.i386.rpm

on a RedHat 9 machine.

The log from the installation has been attached.

The only major error is that the cciss module
isn't updated.

I've run "hpasm activate".

However, if I try to connect to
https://servername:2381, I can log in, but I
get a lot of messages like: servername received a message with incorrect Message Authentication Code. If the error occurs frequently, contact the website administrator. The page appears
after a while, but there are no items
underneath the "Failed & Degraded Items" box.

What have I done wrong? How do I fix it?

[KristofH]

in my opinion the problems arise because you didn't install the adapted snmp packages yet:

ucd-snmp-4.2.5-7cmaX.76.Redhat8_0.i386.rpm
ucd-snmp-devel-4.2.5-7cmaX.76.Redhat8_0.i386.rpm
ucd-snmp-utils-4.2.5-7cmaX.76.Redhat8_0.i386.rpm

But, when trying to install these packages (maybe first remove your current snmp-net packages) you get some errors (will check later today wich one) and you can proceed installing but still snmp won't function correctly. (and result, can't use agents correctly)

please keep us posted on your issues, I'm having the same problem. (and am sure many of us.. )

g'luck!

[Mogens Kjaer · Accepted solution]

I managed to get a few minutes of success:

If you install the ucd-snmp packages, they
will complain of missing libelf.so.0 and
librpm*-4.1.so libraries.

I took the files:

libelf.so.0
libelf.so.0.8.2
librpm-4.1.so
librpmio-4.1.so
librpmdb-4.1.so

from /usr/lib of a rh80 installation and
placed on the rh9 machine, and I installed
the ucd-snmp rpms with --nodeps.

It runs! But after a few minutes, the snmpd
program stops:

$ gdb /usr/sbin/snmpd
(gdb) run -f
Starting program: /usr/sbin/snmpd -f
[New Thread 1077872384 (LWP 3517)]
[New Thread 1086266672 (LWP 3524)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1077872384 (LWP 3517)]
0x4002bf9f in search_subtree_vars () from /usr/lib/libucdagent.so.0

[Mogens Kjaer]

I've tried rebuilding the ucd-snmp rpms
from the src.rpm, under RedHat 9, but
it still fails (having removed the
libelf and librpm symlinks).

I test it with:

snmpwalk -v 1 localhost public enterprises.compaq

and the snmpd process dies after a few
tries. Hm...

[Mogens Kjaer]

I have to give up on this one for now...

I have two almost identical ML370 machines,
one installed with rh8.0, and one with rh9
(eventually both will run rh9 in production).

snmpwalk -v 1 localhost public enterprises.compaq

works without problems on the rh8.0 machine,
on the rh9 it will crash snmpd after a short
while. If I repeat this, it doesn't necessarily
stop at the same place in the MIB.

I've tried compiling ucd-snmp with -O0 instead
of -O2, no difference.

So if this is a memory overflow/timing problem,
I give up :-(

Mogens