Templates for HP Servers

Cova · Joined: 27 Nov 2006 Posts: 15

As I said a while back over in the other HP-related thread - I've been working on a set of templates based on the SNMP information made available by HP's Insight Agents. Some of these OIDs are available on any system with the HP agents installed, some of them are specific to the windows agent. There is a OID tree specific to linux agents as well, however they don't really provide much on an ESX host and I don't have physical HP linux boxes to develop templates against. OIDs that should be available on any OS with the agents installed I've labeled as "generic OIDs".

In the attached zip is a single export of my templates from Cacti 0.8.6j - hopefully it detected all the dependancies correctly and isn't missing anything and also doesn't include too much extra. There are also 6 data queries (the 6 small files) that need to be installed into your <cacti_path>/resource/snmp_queries/ directory.

These templates currently allow for monitoring:
CPU Utilization (User/System time - uses Windows agent OIDs)
Memory Faults (soft/hard faults per second - uses Windows agent OIDs)
NIC Utilization (Bytes/Packets sent/received, queue, errors - uses Windows agent OIDs)
Disk Utilization (Bytes/IOs read/write per second, queue length - uses Windows agent OIDs)
Temperature sensors (current temp + threshold - uses generic agent OIDs)
SmartArray Controller utilization (commands / sec, raid CPU utilization, command latency - uses generic agent OIDs)
SmartArray Logical Drive utilization (read/write IOs per second, cache hits/misses/depth, read/read-ahead/write sectors per second - uses generic agent OIDs)

Update:
There might be some un-documented fixes scattered around. I re-exported it to get the new graphs below included. You're also likely getting my 5-year RRA - since all RRA's are selected by default you'll have to go through and delete that one if you don't want to keep data for that long.

SmartArray Physical Disk info (read/write sectors per second, some static drive info (serial no, firmware, etc.) - uses generic agent OIDs)

YoMarK · Posted: Mon Mar 26, 2007 4:55 am Post subject:

Thank you!
Implementing them right now one some Vmware ESX servers, and they seems to be working great(the non-windows parts).

jkstraw · Joined: 01 Aug 2006 Posts: 34

Great work - thank you for posting these!

apperrault · Posted: Wed Apr 25, 2007 3:17 pm Post subject:

Thanks for these great templates. I am having a problem on a couple of my machines. Some of the data just does not show up on them. At first, i thought it was because the systems are DL360 G5s, and things may have changed from the G4 to the G5, but that doesnt appear to be the problem. On one system, the LogicalDrive IOs doesn't show a graph, but on another one, that graph shows up great, but the ArrayController doesn't seem to work. In the log file for these it says:

Cova · Joined: 27 Nov 2006 Posts: 15

I'm afraid I often have the same problem with some hosts. It seems to be a bug in the HP SNMP agents - rebooting the host usually fixes it. Unfortunately having cacti problems isn't enough justification to reboot production servers here so I usually have to wait till the next microsoft patch cycle.

If you run an snmpwalk against the servers showing that problem you will notice that they don't return any data at all. I've got a windows 2000 / SQL 2000 server here right now that the entire .1.3.6.1.4.1.232.19 tree is missing - though it was there a couple weeks ago. Neither restarting the SNMP service or the HP Insight Agents resolves the problem.

RHarmer · Joined: 24 May 2007 Posts: 1

I'm very curious what the temperature graph looks like. Can someone post some pictures of some temperature data?

I don't have Cacti currently running (I will shortly), but I'd love to see what it can do against HP temperature information.

Thanks!

Rich

kohalik_kalamees · Joined: 23 May 2007 Posts: 2

Good morning fellow mates!
Has anybody made some templates for Disk Array Physical Drives too?
if so, then i would appreciate that a lot.

Cheers.

ajohnson · Joined: 27 Dec 2005 Posts: 36

Note: I had to increase the width of the graph templates to be 525. It was cutting of the "m" in the maximum fields.

Looking at the graphs this morning and I'm seeing 94.68 for maximum command latency and I'm thinking HOLY SH!T! It's taking over a minute and a half to complete SCSI commands!!!

Cova · Joined: 27 Nov 2006 Posts: 15

Ok - did some more work on these. I'll update the original post in a moment with a new version of the attachment. Here's some screens of graphs - this is NOT a complete set of all the graphs included in the host template.

Yes - it now does disk array physical drives too. Anything else you'd like graphed from the HP Insight Agent MIBs, post in this thread and I'll see if it can be done.

apperrault · Posted: Wed Jun 27, 2007 11:07 am Post subject:

These templates ROCK!!!!! I mean they were great before, but the fact that we can now see down to the individual disk level, that is very cool. Now i need to add this to all of my systems. I will look through everything else that HPInsight collects, and let you know if i see anything else useful.

thanks

app

cyrill · Posted: Fri Feb 01, 2008 10:03 am Post subject:

hello !

I've get the seven small file in <cacti-path>/ressource/snmp_queries and after this, I've imported the big file into cacti.

I've created new host and associed the host template. All Associated data Querie are the status "Success" but each of them are 0 items and 0 rows ...

why ?

fmangeant · Posted: Fri Feb 01, 2008 10:06 am Post subject:

Hi

can you run the data queries in verbose mode ?

cyrill · Posted: Mon Feb 04, 2008 1:41 am Post subject:

Hello !

Yes I can run all data query in verbose mode ! But For each query No SNMP data returned...

BSOD2600 · Posted: Mon Feb 04, 2008 2:10 am Post subject:

Post a snmpwalk of .1.3.6.1.4.1.232.19.2.3.2.1

cyrill · Posted: Mon Feb 04, 2008 2:23 am Post subject: