How to scan a farm

[Moe_5]

Hey guys,
im trying to scan our farm of 200+ servers to determine if there are any drive/power supply failures out there. Is there anyway I can run a scan that would give me real time info on the status of these components using CIM or the management agents?

Thanks!
Moe

[Ken Henault]

I'm assuming you've installed the Insight Management Agents on the servers. If you have one of the HP management tools running (CIM, IM7 or SIM), you'll see systems with degraded components as a change in color status. The three tools represent different generations of the same tool. CIM is obsolete and IM7 is just starting to replace IM7. I'd recomend you set up a SIM server.

More information, and the download are available at: http://h18013.www1.hp.com/products/servers/management/hpsim/index.html

[Moe_5]

Here is the thing though, I have 2000+ servers and drilling down each one to see if they have any failed components is not an option, I want to create a script to scan the servers, maybe failed component information is stored in the registry or something (im not sure). Thats what im trying to figure out though, how can I run a realtime scan of my farm? WIthout checking every single one.

[James Kennedy_4]

You don't have to drill down on each server. Just drill down on the ones that come up as Major, Minor, or Critical failures. If the server is green, that means all is well with that hardware.

[Moe_5]

That would mean about 600-700 servers...

[Ken Henault]

With SIM or IM7 you can write a query or report to drill down on this type of information.

[Moe_5]

I have IM7 and I dont see this option, can you please tell me how this is done? And is this realtime?

[David Claypool]

Moe:

Your setup is exactly what Insight Manager (most recently hp Systems Insight Manager) is designed to do. As others have indicated, the first step in the process is to make sure all of the systems you want to manage have the ProLiant Support Pack that includes the appropriate drivers and agents. Once you know where hpSIM is going to be installed, you will want to configure the SNMP service so that it sends events back to the hpSIM application. That way, anytime something happens (fan failure, drive problem, etc) a message will be sent back to hpSIM with detailed information about the problem. This is kind of like calling for help.

In addition, hpSIM can poll all of your servers at regular intervals and ask if they have a problem. The result of that poll is going to be a change in status without any detail, so ideally the agent will already have communicated the problem. However, this polling is a useful redundant method of maintaining information about the health of a system.

[Rob Buxton]

Just stepping back a bit, which column in IM is not showing Green, is it the HW column. If you've got 600-700 show non-green then you'll have some work to do!

Alas, it can be something as simple as uncleared entries in the Integrated Management log, or secondary active NICs with no network connection (or loopback connector).

I don't know of an easy report you can run to get the overall status of say Array Controllers, but the report for Power Supplies does indicate the status and may provide what you want.

To get the best out of IM you'll want to look at clearing errors so that you get only a few Servers showing issues, then they can be managed.
Here, chief amongst those was getting the redundant NICs sorted, either by using loopback connectors or disabling them.