After reboot, managed node shows up as Serial number for it's name

[NJK-Work]

Hi,

I posted about this earlier but closed the post because I thought I had it fixed. I do not.

I upgraded to SIM 5.2 from 5.1. Now, every time a managed node is rebooted, SIM changes its name from the DNS/host name to the serial number of the server. It also locks the managed server down in SIM as "In Maintenance" and will not let be bring it out of maintenance. The only solution is to delete the object and manually re-discover it.

Setting the system properties to "Not change name" does not work. Upgrading to SP 7.91 I believe does fix the problem, but I really cannot upgrade 200+ physical machines (virtual machines are not doing this I believe) to SP 7.91 right now - that is just too many outages.

If anyone can provides some tips, I would greatly appreciate it.

Thanks
Nelson

[NJK-Work]

After some testing, I have found that restarting SNMP on the problem server (which then restarts all the HP agents) magicially fixes the problem.

So my theory is that when the HP agents (or SNMP) is restarted - in this case by a reboot of the server, it triggers something in SIM causing it to lose it's identity and change it's name from HOST to Serial number. Then, restarting the agents a second time (while still in a bad state in SIM) reverses the problem and all is back to normal. This can be done either manually by restarting the services or a reboot.

The problem from that point on seems to go away for that particual server.

One thing I have enabled is "Automatically discover a system when an event is received from it" under "Discovery/General Settings". So the managed node is sending a SNMP trap to SIM when it restarts and SIM must be "rediscovering" it incorrectly.

Anyone else have this option enabled (and NOT having this problem)?

Thanks
Nelson

[Rob Buxton]

I've installed 5.2 on a new virtual server. At this point it doesn't have snmp access to the servers. I've noticed a couple of servers have entries with server names. These are our VMHost Servers and in both cases have ILOs attached.

It looks as there's some issue with how servers are discovered via the ILO. Not sure if it's quite the same issue you have.

[NJK-Work]

Interesting...I will investigate the iLOs.

Thanks
Nelson

[Rob Buxton]

I meant servers identified with serial numbers above as you're seeing.