EMS and a VA7410

[Don Mallory]

Here's the setup:

I have 4 boxes (that I am concerned about at the moment). Two rp54xx's (pike2, pike5), an rp2xxx (pike4) and a K570 (pike1). All are running HP-UX 11.11. 3 of the units were installed with the June 2002 Support Tools CD's, one of the rp54xx (pike5) was installed from the March 2003 CD. All four boxes have commandview installed (/opt/sanmgr/commandview), all four are connected to a VA 7410 via A6685A TachLite Fibre Channel card.

EMS Versions are A.03.20.01 on all of the boxes.

PHSS_27850 - OV EMANATE14.2 snmpdm - obsolete mib patch is in place on all boxes.

EMS is configured to send all events of

>= MAJOR to SNMP, SYSLOG and /var/opt/resmon/log/event.log
= MINOR to SNMP and /var/opt/resmon/log/event.log
= INFORMATION to SNMP and /var/opt/resmon/log/event.log

When everything is working, I will stop sending INFORMATION and MINOR to SNMP.

snmpd.conf is set with a non-public community name (as is standard for our entire site), and the trap destination is set.

When I started:

pike1 and pike4 did not have SNMP configured to our standard community name or any trap destinations set, pike2 and pike5 wer configured to our standard.

Sanmanager was not installed on pike4

Reconfigured all the traps on all four boxes as described above.

ENABLED DISK_ARRAY, DISK_DRIVE, JBOD and FC_SWITCH in the /var/stm/config/tools/monitor/RemoteMonitor.cfg file.

None of the boxes have been rebooted lately (since I started trying to get EMS working) due to application deployment work which is being done very heavily on them at the moment. I'd just like this to work before they go production.

Here's the problem:

pike5 is sending traps through to NNM via SNMP, all of the other boxes are configured identically and are not sending traps to NNM via SNMP.

All four boxes are unable to send traps related to the attached VA 7410 either to the event.log or to SNMP.

A send_test_trap -a -v disk_em works on pike5 to NNM, and to the event.log on all other boxes.
A send_test_trap -a -v RemoteMonitor does not work on any box to event.log or SNMP.

I was working with a tech at HP support who sent me a program he wrote in C to force a test trap to the RemoteMonitor, this however does not answer the question of why does the standard tool not work?

I have stopped and started the /sbin/init.d/diagnotics services, all of the Snmp* services, ems, and have killed the p_client allowing it to restart. ./diagnostics stop only kills the diagmond, all of the other processes have to be killed by a kill -SIGINT PID manually (or by a for item in `ps -ef | grep stm | awk '{print $2}'`... kill $item...)

I have restarted all the services using the exact order as they are in /sbin/rc0.d through rc3.d.

So, that's where I am. Any suggestions?

Thanks,
Don

[Cheryl Griffin]

Don,
Due to the scope of this issue, I believe you should still be working with HP Support on this. Given what you have posted here, the first thing I would check is your versions of everyting involved here: OV, SNMP, Diagnostics, EMS, Patches, etc.
Cheryl

[Don Mallory]

I appreciate your suggestion, however I have been working with an HP Support Technician for 10 days now and have been bounced around with little help beyond where I started.

I am looking for other suggestions from anyone who may have seen a similar issue.

Thanks,
Don

[Don Mallory]

I have attached a zip with the swlist and swlist -l fileset from "pike1".

Quick answers:

SNMP - B.11.11 or "SNMP Agent Resident Module Version 14.2.1.7"

OV - doesn't matter, it works for every other server on the network, and is receiving traps from one of the 4 units ("pike5").

OnLine Diags - B.11.11.07.11

EMS - A.03.20.01

Thanks again.