SIM 4.2 and ILO

[Steve Giordano]

I have a couple of servers where the ILO is activated but the "MP" column in SIM is still showing a "?". The snmp settings are identical on the SIM server and the managed server.

Also the system information is not being populated when a server is dicovered. Again all snmp settings are identical. Thanks for any help.

[Ed Cox]

Steve,
On those servers where you are seeing the ? in the MP column... go to one of them and check under control panel - HP Management Agents applet and check to see if your Remote Insight Information agent is on the Active list. If not ...activate the agent - you will then be prompted to restart the agents.
Then go back to SIM and do a Daily Device discovery...

As far as the second issue:
What do you show under the Autodiscovery - Discovery Filters area for the default SNMP community strings? Do these default fields match up with the strings on the new servers being discovered?
Ed Cox

[David Claypool]

Ed's only partly correct. While it is a good thing to make sure that the 'Remote Insight' component is active in the Insight agent control panel applet, it has nothing to do with iLO discovery.

First, the iLO itself has to be a discovered device on its own. Check the 'All management processors' list to see if it shows up. The association then happens by retrieving information from the iLO.

During initial or daily device identification, HP SIM asks each iLO some information about itself to which it responds with an XML document. You can view that information from an iLO by typing

http://[ILO-NAME-OR-IP-ADDRESS]/xmldata?item=all

to which you will get a response such as this:




6J2AJNT1B02G


1
Integrated Lights-Out (iLO)
ILO6J2AJNT1B02G



The level of detail that is returned can be set under Administration-->SNMP/Insight Manager Settings. The basic data used for the association is the server's serial number, tagged by above. On a discovered iLO, HP SIM searches the database for a server with the serial number it has received from iLO in order to make the association.

Anything that causes a bad serial number to appear (such as after a motherboard replacement) will interfere with identification.

[Steve Giordano]

I did run a discovery on our management VLAN and it discovered all "servername-ilo" connections so that's working now. The one "Server" type that was discovered and has all system information made the correct association probably (like you said) the ILO was discovered as well. I'm still trying to get the System information to show up on other test servers. If I discover manually it will show up as "Unknown" device. I'm sure if I can get around that problem I'll have my ILO association problem fixed as well.

[David Claypool]

I think this will be a great application of the "Configure or repair Windows agent" feature new in 4.2. It will go out and fix any SNMP problem and create the trust relationship for you. It will prompt for Windows credentials through which it will make the connections to the remote boxes.

[Steve Giordano]

I'm getting authentication failed when I run the Configure/Repair Windows agent. I'm an admin on the server too. Weird. I'm able to discover servers now without any problem. But the "Management Processor" system type is not syncing with the "Server" system type except for one server. That one server has a green check in the "MP" column but not the others and the ILO's are discovered properly on all servers. I can't see anything different on this server than the others. The SNMP settings are identical and the Management Processor System List even found "servername-ilo on server servername."

[Joel Rubenstein]

The "repair agent settings task" requires that openSSH be configured on the CMS. It runs the task as the "Administrator" user.

A quick test to determine if openSSH is setup correctly is to select the CMS, and then go to options --> command line tools --> windows --> netstat. If this fails then openSSH is not properly configured.

[Jon Ward]

Did you check the iLO Insight Manager settings within it's web management interface?

[Steve Giordano]

The frustrating part of this is now the server that was showing the check mark in the "MP" column is now showing a "?". I just setup the SNMP settings for SIM in the ILO interface. It's still detecting all ILO's under the "All Management Processor" system list but it won't associate it with the same servers in the "All Servers" list. One server did work (it was showing the check mark in the "MP" column -- I'm not sure how/why) but now it's showing the "?".

[Jon Ward]

Do the associated Management Processors have green status marks? Is their names "iloname in Server servername". This is an indication of the proper operation of the iLO itself.

Something a simple as a rerun of the Autodiscovery on all relevant ip addresses may correct the situation if you want to hurry up the process.

[Steve Giordano]

Yes -- the Management Processor system list is showing "servername-ilo on server " That's why I'm so stumped as to why it won't make the association in the All Servers list. I think I'll try the Autdiscovery again too.

[Steve Giordano]

Deleting the server and rediscovering worked. The association was made in the "All Servers" list and I can access the ILO from there in the "MP" column. I'm going to try a couple more test servers and see if that helps.

[Steve Giordano]

I was able to add 3 more servers successfully. The "MP" column has the check mark and the ILO is accessible through Insight Manager. BUT -- I refreshed a few hours later and now 2 of the 3 servers have the "?" in the MP column. UGH!!! All information is still present except for the check mark. Is there some kind of job/stored procedure that is looking for a flag that keeps getting updated or something of that nature?

[Jon Ward]

The only thing I can think of doing is to make certain everything is upto date. The most important item to ensure is current is the iLO firmware. The current firmware is 1.64 and is available from http://h18000.www1.hp.com/support/files/server/us/download/22156.html

[Steve Giordano]

I scheduled an Automatic Discovery every 2 hours and that seems to work. Functionality returns after a re-discovery. I wasn't doing any auto discovery since I'm only in a test "mode" for now but this is working. Thanks for all the responses -- I appreciate everyone's help.

[Darrin Rawls]

Running automatic discovery that often should not be the 'fix'. Please log a support case so that this can be escalated and evaluated.

thanks!

[Christopher Maddux]

Well, it's been a month and half. Any one have a solution to this problem yet?

I am having the same problem here. While I look at each individual server or ILO, they show linked up. But the MP column shows ?

Christopher

[Steve Giordano]

I'm in the process now of rolling this out to production because we NEED HW monitoring in place and everything was going well the first few days. I had 23 servers discovered with all options showing up correctly and just today the problem with the MP column going from check the "?" is back. Could this be a problem with SQL permissions? Sounds odd I know but I initially had a problem where I couldn't see any server I discovered until we gave the service account DDLAdmin rights in SQL. We tried using just DataReader/DataWriter and that didn't work. Some EtherReal traces found permission errors running certain operations against the database. I'm really reaching now because this has been an ongoing problem. Does anyone have the support number I can call if we can't get this resolved through the message board? I'd like to use that as a last resort.

[Jon Ward]

If you have warranty, goto http://www1.itrc.hp.com/service/mcm/countrySelection.do to log a case. If you have a contract, goto http://www1.itrc.hp.com/service/mcm/homepageRequest.do .

Alternatively, you may call (800) HP-INVENT, then say "Insight Manager".

They will ask that you use the iLO firmware 1.64 and hp Systems Insight Manager to ensure that the hardware/software combination is supported.

[Christopher Maddux]

I think the best alternative is to revert back to v4.1 when it worked. Unfortunately v4.2 fixed several other things.

I won't bother calling. I can't remember the last time they actually fixed something other than a failed HD. And even that's becoming difficult.

Christopher

[Darrin Rawls]

If you call support, on the voice recognition, say "Insight Manager". The call then gets routed to DEDICATED support folks for HPSIM/plugins/blades. They are trained on the product(s) very well by the way....

-Darrin

[Preston Gallwas]

Heh...tried calling,

"The average wait time in the last thirty minutes was ... fourty two minutes..."

(counts on fingers)

How does that work? hehe

[Christopher Maddux]

Well after waiting for 28 minutes on hold (better than 42 minutes...), I finaly got a live person on the phone. And of course, this guy has never heard of this problem before. This is the typical response I get every time I call.

After about 10 minutes of mostly silence on the other end with the occasional question, I am back on hold...

I would think that if someone has called HP on this before, they would have a record of it.

[Darrin Rawls]

Youch; will have to look into that one....

[Christopher Maddux]

Well as expected. Another 20 minutes on hold just to have the guy come back and tell me that he can't find anything on it and his is working fine.

I have a case number on this now, but I have several cases that have been open for months with no updates. So I guess that's the end of that.

That was an hour of my life I will never get back...