CPU/Disk utilization through snmp?

[Sean OB_1]

Howdy,

Does anyone know if it's possible to get CPU/Mem/Disk utilization through the Emanate SNMP agent on HPUX servers?

TIA and points for all responses.

Sean

[Steven E. Protter]

I believe it is.

My toptools admin use to be very proud of himself when he could tell me my HP-9000 box was redlined on cpu before I noticed it (like I didn't have better things to do).

Its also possible to set up the MIBS to prevent this. SNMP can be a security hazard. Some of its information is not transmitted secure, especially in older releases.

SEP

[Sean OB_1]

Anyone know how this can be done? By default all you can see are the nics as far as I can tell.

Ideally this client would like to see CPU, Memory and disk capacity through snmp queries.

[Mel Burslan]

Sean,

I have never done this myself but I think SNMP by itself does not provide this functionality but EMS does and EMS subsystem has an agent which talks SNMP.

man emsagent can tell you much more than I can as you seem to be more knowledgable than me as it comes to SNMP.

HTH

[Dave Hutton]

You can it looks like. Just for the heck of it I would do this:
servera:root /root # snmpwalk servera | grep -i mem
hp.nm.system.general.computerSystem.computerSystemFreeMemory.0 : Gauge32: 2787732
hp.nm.system.general.computerSystem.computerSystemPhysMemory.0 : INTEGER: 7340032
hp.nm.system.general.computerSystem.computerSystemMaxUserMem.0 : Gauge32: 6960900

I'm guessing the other values are there too for at least cpu. Disk maybe tough.

And obviously you can go against a specific snmp attribute:
servera:root /root # snmpget servera hp.nm.system.general.computerSystem.computerSystemFreeMemory.0
hp.nm.system.general.computerSystem.computerSystemFreeMemory.0 : Gauge32: 2764688

[Dave Hutton]

No need for points on this. Just noticed you can get all kinds of information, hard to know what the values mean. Like looking at cpu:
servera:root /root # snmpwalk servera hp.nm.system.general
hp.nm.system.general.computerSystem.computerSystemUpTime.0 : Timeticks: (122813000) 14 days, 5:08:50.00
hp.nm.system.general.computerSystem.computerSystemUsers.0 : Gauge32: 8
hp.nm.system.general.computerSystem.computerSystemAvgJobs1.0 : Gauge32: 55
hp.nm.system.general.computerSystem.computerSystemAvgJobs5.0 : Gauge32: 61
hp.nm.system.general.computerSystem.computerSystemAvgJobs15.0 : Gauge32: 52
hp.nm.system.general.computerSystem.computerSystemMaxProc.0 : INTEGER: 4100
hp.nm.system.general.computerSystem.computerSystemFreeMemory.0 : Gauge32: 2788880
hp.nm.system.general.computerSystem.computerSystemPhysMemory.0 : INTEGER: 7340032
hp.nm.system.general.computerSystem.computerSystemMaxUserMem.0 : Gauge32: 6960900
hp.nm.system.general.computerSystem.computerSystemSwapConfig.0 : INTEGER: 9437184
hp.nm.system.general.computerSystem.computerSystemEnabledSwap.0 : Gauge32: 9437184
hp.nm.system.general.computerSystem.computerSystemFreeSwap.0 : Gauge32: 9437184
hp.nm.system.general.computerSystem.computerSystemUserCPU.0 : Counter: 13601386
hp.nm.system.general.computerSystem.computerSystemSysCPU.0 : Counter: 18729642
hp.nm.system.general.computerSystem.computerSystemIdleCPU.0 : Counter: 79838902
hp.nm.system.general.computerSystem.computerSystemNiceCPU.0 : Counter: 259685
hp.nm.system.general.fileSystem.fileSystemMounted.0 : Gauge32: 13


Hopefully thats what your looking for. To be able to query a server for specific information.

[Devesh Pant_1]

Why snmp only ? there are many other better ways

[vinod_25]

hi

To do this two collections will need to be configured, one with the threshold set at 30% and the second with the threshold set at 80%. Since it is not possible to set up two collections on the same mib expression it will be necessary to edit

$OV_CONF/mibExpr.conf

Add two copies of the If%util entry, each with a different name (for example If30%Util and If80%Util).


** Warning **

Before editing mibExpr.conf make sure you have saved a copy of it since an incorrect mibExpr.conf file could seriously impair the functionality of NNM.

** **


Copy the If%util entry and paste it in twice, remembering to separate each entry by a row of hash (#) symbols. Rename one of the new entries to If30%util and the other to If80%util. Save the mibExpr.conf file then stop and restart snmpCollect.

From the [Options] menu in the NNM window choose [SNMP Data Collection and Threshold]. In the dialogue box that opens up click on the [Edit] menu and select [Add] then choose MIB Object. A list of MIB objects appears and the newly added expressions should be included in that list.

Configure a collection for each of these two expressions, setting the appropriate threshold for each collection. Configure the threshold event to send the required message when that threshold is reached.

Bear in mind that each collection increases the amount of traffic on the network.

regards

Vinod K

[Sean OB_1]

"Why snmp only?"

Because the tool they are using only uses SNMP.

They do not have NNM/OVO nor do they have any plans on purchasing it.

[Mahesh Kumar Malik]

Hi

Please visit following URLs:

http://www.net-snmp.org/FAQ.html

http://www.net-snmp.org/#Documentation

Mahesh