|
|
| Author |
Message |
helpmeplease
Beginner
Joined: 19 Nov 2004
Posts: 3
|
 Posted: Fri Nov 19, 2004 10:58 pm Post subject: CPU graphs out of scale |
 |
|
| I'm showing CPU util data graphing at over 100%. I didn't have this problem previously using RRDTool by itself, updated by a script containing snmpget commands. Does the counter data that it is retrieving differ from what I was getting before? Previously, I was doing an snmpget on ssCpuRawUser.0 for example, and I noticed that 'Linux Box' template uses 'enterprises.ucdavis.systemStats.ssCpuRawUser.0'. Is there any difference between these two that might account for the difference in my graphs? If so, should I create a new test and change everything to use it, or is there a way I can manipulate the counter data that is being retrieved to reflect the correct percentages in the graph? I did a search on this and couldn't find any useful information about it. |
|
| Back to top |
|
 |
silfreed
Site Admin
Joined: 19 Jan 2003
Posts: 394
Location: Shippensburg, PA, USA
|
| Posted: Fri Nov 19, 2004 11:38 pm Post subject: |
|
|
Those OIDs should be the same. Commonly we do see values over 100% for multi-processor boxes; net-snmp doesn't handle this at all internally. Is this the setup you have?
-Doug |
|
| Back to top |
|
|
helpmeplease
Beginner
Joined: 19 Nov 2004
Posts: 3
|
| Posted: Sat Nov 20, 2004 6:58 am Post subject: |
|
|
| Yes, I'm sorry, I forgot to mention that. These are dual-processor boxes, WhiteBox 3.0 (2.4.21), net-snmp 5.0.9. The only reason I asked about it is because I didn't see the problem when using the snmpget command launched from a shell script. What is different could possibly account for the difference? |
|
| Back to top |
|
|
balleman
Site Admin
Joined: 20 Jan 2003
Posts: 282
|
| Posted: Sat Nov 20, 2004 3:34 pm Post subject: |
|
|
| Maybe the shell script looks for the number of processors and scales accordingly? If you want, you could create a dual-processor CPU template graph (just duplicate the one provided) and set multipliers on the items for 0.5. We normally just graph the values as provided, keeping in mind that a 2-processor box can use 200% CPU. |
|
| Back to top |
|
|
helpmeplease
Beginner
Joined: 19 Nov 2004
Posts: 3
|
| Posted: Tue Nov 23, 2004 2:48 am Post subject: |
|
|
| balleman wrote: |
| Maybe the shell script looks for the number of processors and scales accordingly? If you want, you could create a dual-processor CPU template graph (just duplicate the one provided) and set multipliers on the items for 0.5. We normally just graph the values as provided, keeping in mind that a 2-processor box can use 200% CPU. |
Yeah, I'm thinking that's what I'm going to do. I was just wondering if there were something else wrong, as I wasn't sure if it was reporting double the CPU usage for all of my boxes, even the single-processor ones. Since it doesn't look like that's the case, I'm just going to custom graph these few boxes. Thanks for the help. |
|
| Back to top |
|
|
|
