What is this?

This knowledgebase contains questions and answers about PRTG Network Monitor and network monitoring in general. You are invited to get involved by asking and answering questions!

Learn more

PRTG Network Monitor

Intuitive to Use. Easy to manage.
300.000 administrators have chosen PRTG to monitor their network. Find out how you can reduce cost, increase QoS and ease planning, as well.

Free Download

Top Tags


View all Tags

IBM Power monitor

Votes:

0

Your Vote:

Up

Down

Hi Paessler
We are in the process of installing several IBM Power servers. When running Linux on these, linux sees all CPUs in the physical power server, even though it only has access to a few of them. This makes the monitoring of my servers on Power "not valid". I ask of you to consider looking into a way we can monitor resources properly on the IBM Power platform.

cpu cpu-utilization prtg

Created on Mar 24, 2021 7:56:36 AM by  Jacobnm (10) 1

Last change on Mar 24, 2021 12:30:40 PM by  Felix Wiesneth [Paessler Support]



9 Replies

Votes:

0

Your Vote:

Up

Down

Hi Jacob,

I'm afraid that we do not have experience with monitoring of IBM Power Monitor we could share with you. Does one of these sensors work?

If not, you can see if there is some kind of API you can use to gather data, or if IBM provides an SNMP MIB file, which could be integrated by using our MIB Importer Tool and create an SNMP Library Sensor afterwards.


Kind regards,
Felix Saure, Tech Support Team

Created on Mar 24, 2021 12:42:51 PM by  Felix Saure [Paessler Support]



Votes:

0

Your Vote:

Up

Down

Hi Felix

Thanks for your answer. I am sorry, but we can not query the Power servers in the same way as we do it regular x86 systems. They do not have an ILO or iDrac interface in the same way. They are all managed via a management server. Which can only do SNMP traps and Email.

Best regards Jacob

Created on Mar 24, 2021 2:52:04 PM by  Jacobnm (10) 1



Votes:

0

Your Vote:

Up

Down

Hi Jacob,

If no active monitoring like SNMP or HTTPS is allowed, you can use the passive receivers:


Kind regards,
Felix Saure, Tech Support Team

Created on Mar 25, 2021 4:30:11 AM by  Felix Saure [Paessler Support]



Votes:

0

Your Vote:

Up

Down

Hi Felix

Thanks for your answer. That is true, and we are already using SNMP Traps to monitor the hardware itself. What I am looking after is the monitoring of the virtual machines running on the Power system. We have installed several Linux installations on the power systems. It is primarily the CPU utilization that is the problem. It can not remove the cores not available to the machine, like the service "https://lpar2rrd.com/" can do.

Br Jacob

Created on Mar 25, 2021 7:45:01 AM by  Jacobnm (10) 1



Votes:

0

Your Vote:

Up

Down

Hi Jacob,

Indeed, the sensor performs a meta scan to discover all available CPUs. Are you using the SNMP or the SSH CPU sensors?

For the SNMP, you can put some manual afford in creating the sensors. Perform an SNMP Walk to get a list of all available OIDs pointing to the CPU. You can walk the OID

1.3.6.1.2.1.25.3.3.1.2.index

for that, the last digit (.index) is the identifier of the CPU then. This way, you can add an SNMP Custom (Advanced) Sensor to the device and manually add the CPUs.

Otherwise, a custom (SSH) script might be the way to go if you wan to do active monitoring.


Kind regards,
Felix Saure, Tech Support Team

Created on Mar 25, 2021 8:30:31 AM by  Felix Saure [Paessler Support]



Votes:

0

Your Vote:

Up

Down

Hi Felix

We are currently using SNMP. The adjustments can possibly be made yes. Just need to identify which CPUs it can use. I will work and see if they can be identified in the OS. But it still does not change the wish of feature request to support Power :)

Br Jacob

Created on Mar 25, 2021 11:51:37 AM by  Jacobnm (10) 1



Votes:

0

Your Vote:

Up

Down

I have been talking to a technician about this. This is the output from one of our servers:*Architecture:

ppc64le
Byte Order:          Little Endian
CPU(s):              32
On-line CPU(s) list: 0-31
Thread(s) per core:  8
Core(s) per socket:  1
Socket(s):           4
NUMA node(s):        1
Model:               1.2 (pvr 004e 2102)
Model name:          POWER9 (architected), altivec supported
Hypervisor vendor:   pHyp
Virtualization type: para
L1d cache:           32K
L1i cache:           32K
NUMA node0 CPU(s):   0-31
Physical sockets:    4
Physical chips:      1
Physical cores/chip: 10

In nmon I can see that the "real" number we want to see is called "PhysicalCPUused" in nmon - that actually shows the right performance number. Can that help in any way?

br Jacob

Created on Mar 25, 2021 12:47:15 PM by  Jacobnm (10) 1

Last change on Mar 25, 2021 2:34:10 PM by  Felix Wiesneth [Paessler Support]



Votes:

0

Your Vote:

Up

Down

Hello Jacob,

Then I'd recommend to write your own script to extract the desired metric and to forward it to PRTG. This can be either actively via the SSH Script Advanced Sensor or by pushing the script's results to PRTG using the HTTP Push Data Advanced Sensor.

We have a couple of sample scripts in our PRTG Sensor Hub to get an idea about the required syntax - none with this special requirements so far.

It would also be great if you would upload the script afterwards for the community.


Kind regards,
Felix Saure, Tech Support Team

Created on Mar 26, 2021 4:56:13 AM by  Felix Saure [Paessler Support]



Votes:

0

Your Vote:

Up

Down

Hi Felix

Thanks for your recommendation. I have not found a command to query it yet, but if I find it, I will add it to a script. Have a nice day.

Best regards Jacob

Created on Mar 26, 2021 7:22:57 AM by  Jacobnm (10) 1



Please log in or register to enter your reply.


Disclaimer: The information in the Paessler Knowledge Base comes without warranty of any kind. Use at your own risk. Before applying any instructions please exercise proper system administrator housekeeping. You must make sure that a proper backup of all your data is available.