Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials

Why does SAM return different hardware between identical IBM servers?

So first, I love the hardware monitoring that SAM 5 brings to the table. It's pretty nice! But one strange thing:

I have two IBM servers, both are 7979-E9U and both have Systems Director 6.3 Common Agent and ServeRAID Manager installed.

The first one shows Fans, Power, Temp, Disk, and Array:

The second shows Fan, Power, Temp, Memory, and CPU:

Anyone have an idea of where the difference comes in? And what I can do to get all the hardware monitored?

Is there some limit of 5 hardware items that can be queried?

Find more posts tagged with

Accepted answers

gryansmith

Updates:

First, after working with support, it looks like the root cause is that the IBM CIM and Adaptec SNMP responders were not replying in time on some polls. This was alleviated somewhat by uninstalling both IBM Common and ServeRAID, then installing ServeRAID 9.3 first, then installing Common Agent 6.3. After the install, it was necessary to restart the Common Agent services a single time to get hardware information to come out.

Second, after reinstalling the software and restarting the services, a piece of hardware will still occassionally disappear (sometimes Memory, sometimes CPU, etc). According to Solar Winds support, this is b/c the poller on our Orion box is timing out waiting for the complete reply. I'm still working on the ticket and the current expectation is that this is a dev fix for a future release.

All comments

Bronx

Take a look at this screenshot. I would verify that this box is checked on your first server.

aLTeReGo

It would seem the 99EE296 node has the necessary ServeRAID management software installed and the other does not. As to why the CPU and Memory information is missing from the 99EE296 node I'm not sure exactly. I've seen this occur with IBM systems running older versions of the IBM Director agent software. I would recommend upgrading to the latest version of IBMs Director agent (6.3 I believe) on that node to see if that resolves the issue.

http://www-03.ibm.com/systems/software/director/downloads/agents.html

gryansmith

First, the CPU & Memory box is checked on both servers.

Also, ServeRAID is installed on both servers. However, the server that reports Disk and Array health has to lower version (9.0). The other server is running 9.2.

Finally, I verified that I installed version 6.3 of the Systems Director Common Agent for Windows.

Perhaps I should downgrade ServeRAID?

aLTeReGo

Perhaps I should downgrade ServeRAID?

That's certainly one thing to try, but let's hold off on that for now. Does this server have an Adaptec or LSI RAID controller installed? Is this a Windows, Linux, or ESX server? How is this node managed? Is it an SNMP, WMI or ICMP node?

gryansmith

It has an Adaptec RAID controller, 8i. And it's running WIndows 2008 R2.

Also, it's an SNMP and WMI node. It's a SQL database server and I use APM templates for monitoring the SQL services (SNMP) and Database Mirroring performance (WMI).

The OS version, node type, network, and everything else i can think of is the same between this machine and the one where disks and arrays are reported correctly.

The only difference I can think of is that this machine was upgraded from Windows 2k3 x64 and the working machine is a bare metal 2008 R2 install.

DONDERKA

Such a difference could be caused if the servers are polled by different polling method for hardware health information's.

For example server 1 is polled as VMWare and server 2 as SNMP.

You can check this on Node Detail page.

Go to the page and locate Node Details resource. Click on Edit Node button.

Check what type of Polling method for this node is selected and if the Poll for VMware check box is checked.

Now we start with the check box. If its checked and the server was added to APM before upgrade to version 5.0 the servers is automaticaly polled as VMware.

If the node is set to polling method SNMP or WMI and the check box is also checked and the node was added after upgrade to SAM 5.0 than the SNMP probably is used for polling for Hardware Health info.

If the servers are set differently here then is high probability that this cause the difference. Because different polling method access the hardware information's through different channels and those channels sometimes give different info.

There is no limitation for how many categories of hardware or sensors you can see.

If the polling method is the same for both servers and they are in all other things similar (or identical) than there could be something wrong with the IBM common agent. How ever I am unable to tell that for sure without diagnostic. If you did open support case could you please post its number? Thanks

gryansmith

I don't believe it's b/c of a difference in polling type. Both servers are setup as SNMP devices and neither have ever been monitored as VMware. In fact, all of my VMware servers are monitoring hardware correctly.

I did open a ticket (#323377) and have been working with support.

Currently the problem appears to be on the IBM side, with certain portions of the MIB tree missing for each server. I suspect this has to do with installation order.

On one server, I install ServeRAID 9.0, then Common Agent 6.3 and got one set of hardware. On the other server, I installed ServeRAID 9.2, then Common Agent 6.3 and got a different set of hardware. Then I updated both servers to ServeRAID 9.3 and got two more different sets of hardware.

The next solution to try is to reinstall Common Agent 6.3. If that doesn't work, then I need to uninstall both items and reinstall, with ServeRAID going first.

I'll post back with the results when I get this completed. Thanks!

gryansmith

Updates: