The SMH on an ESX is partly empty. How to bind hpsmhd to all components.

[rijsbergen]

Hello,

we inherited a couple of ESX V4 hosts for monitoring. Part of them are ok in SIM but some others are not.

On the NON good i checked the following:
- service hp-snmp-agents status
This is the same output as the good one.
All is OK.
So to me all the agent components are ok.
- service hpsmhd status
on the good system it has 11 PID's
on the NON good it has 6 PID's
So to me here seems to be the problem.
- service --status-all is the same on both
ports are opened
- the CMA.conf has no excludes.
the good and NON good are the same.

SIM server
- discovers both
- The certificate was added and i have SSO to both SMH's.
- The NON good system is NOT discovered properly and the SMH only shows some info in 'Storage' and 'System'.

To me it seems that the hpsmhd is not connected to all the components.

Q: How can i 'connect' the hpsmhd to all the components so i have the same amount of PID's.

regards, Marcel

[Taha Tungekar]

Hi Marcel,

Have you tried re-installing the agents? Also, you could compare the version of the ESX agents on the good and the NON good systems.

You could also try using Configure & Repair Agents in HP SIM to deploy the ESX agents.

During discovery / identification, is the output for "good" and "non good" systems the same?

-Taha

[rijsbergen]

>>Have you tried re-installing the agents?
No, re-install means reboot. Don't want to do that yet.

>>Also, you could compare the version of the ESX agents on the good and the NON good systems.
The versions are the same, the kits are too. And as said the 'service hp-snmp-agents status' give the same result. ALL the processes are 'OK' and the same names/amount as on the good systems. Compared side-by-side. So snmp agents seems to be installed properly.

The only difference i see is that the 'service hpsmhd status' process on the good systems have 11 PID's in the list and the non-good 6. So to me it seems some sort of connecting hpsmhd to all the proper processes. I am hoping this can be done without re-installing. Similar to editing CMA.CONF and exclude components (which btw is the default, so no excluded components).

>>You could also try using Configure & Repair Agents in HP SIM to deploy the ESX agents.
Again, this will probably do a reboot.

>>During discovery / identification, is the output for "good" and "non good" systems the same?
No, because it is still an unknown system. The 'hardware description' contains Linux and ESX and more, so the SNMP discovery is OK. But because not all modules are connected to hpsmhd (i assume the cpqhost[d] is the one the discovery gets the system type from) the discovery is not complete. But as said these process ARE running with 'OK'.

regards, Marcel

[SIT]

Hi

I've a lot of ESX server and i've never reboot for a re-install or install .. you can do it !!

During the installation of ESX agent, did you accept all the port in the firewall,did you use a new snmp conf file (with the right community)
Did you instal lthe last agent ?

What is you ESX version ? Agent ?

[AWo]

Try to install the SMH templates. That didi it for me as they hold the GUI content you needd to dsiplay all the stuff.

AWo

[rijsbergen]

@SIT

The README.txt tells me:
....BEGIN....

HP Insight Manager agents have been configured successfully!

HP SNMP agents are installed and configured
Reboot the system to make the changes effective.
------------------------------------------
After above successful configuration message is displayed, reboot the ESX host
with the reboot command.

....END....

So that's why i didn't try.
That message is seen during the installation and we always reboot after this message.

And VM's must be suspended or powered off before installation! See installation doc.

About ports/firewall: see original entry. All is the same.
Version: ok an older one 8.3.1 but it works on other ESX-en, so that shouldn't be an issue.

=============================

@AWo
Do you know perhaps the command for just this part? The vib file is there.
But again I assume that will put the ESX host in maintenance and the VM's must be suspended/poweroff.

I'll see if i can get one of those host for myself and do a re-install.