Aggregation Error and Compatibility Error

[Lapinou]

Hi,

since we upgrade to HP SIM 7.1 update 2 (fresh install), many servers are reporting two errors :

Aggregation Error (Every 5 min)

Event Severity Warning
Cleared Status Cleared
Event Source ex2-bi
Associated System ex2-bi
Associated System Status Normal
Event Time lundi, 19/11/2012, 09:06 CET
Description An aggregation error has been detected
Event Category Power/Thermal Events
Assignee
Comments

Problem Description: System ex2-bi is not a candidate for power history calculation.
Recommended Actions: Refer to the troubleshooting section of the Insight Control power management user guide for more details.

=========================

Compatibility Error

Event Severity Warning
Cleared Status Cleared
Event Source ex2-bi
Associated System ex2-bi
Associated System Status Normal
Event Time lundi, 19/11/2012, 08:56 CET
Description A compatibility error has been detected
Event Category Power/Thermal Events
Assignee
Comments

Problem Description: System is not supported by Insight Control power management
Recommended Actions: If the System is supported by Insight Control power management, and still it appears as not supported re-run identification on this node

===============

Is it possible to disable power management only for systems that seems to be not compatible or something to stop that ?

Thanks,

Julien

[PaulOBrien]

Hi,

Just wondering if you found a solution to this as i am having the same issue with that version of SIM/ICE.

[RCB_2]

Hi:

We also face the same kind of "pollution" problem. In our case, i can´t see the server model's in the Matrix, so i think it's not supported for Insight Power (These servers are Windows, ProLiant DL580 G5)

BTW I think there should be some way to add a exception list or change some collection in order to SIM stop's polluting the events with "aggregation error" events or "A compatibility error has been detected".

Since we updated to 7.1 update 2 we're trying to stop this behaviour, but unfortunately we have not found a solution for exception some servers without disabling the "insight power" feature at all.

Somebody know how to disable single system or a specified fixed collection, or a workaround for stopping this annoying messages?

Thanks in Advance,

BR

[RCB_2]

Ok, i found something that can be used to override the problem.

- Find the server you want to fix. In our case, i did a lookup for: ProLiant DL580 G5

- Open the hostname clicking on it's name

- In "System" tab, navigate down to section: "Power Management", click the "+"

- On the right, select "Configure"

- Select a positive value in the field "Maximum Possible Power Consumption (Watts)" .

I still need help which fixed value should be configured there, i guess that value should be found on the server model's quick specs, e.g. http://h18000.www1.hp.com/products/quickspecs/12770_div/12770_div.PDF 

but i still need to do some research about how to properly configure that value. I configured as a "test-error-test" the value 900 and errors dissappears.

So, the last question will be, how can we guess on HP unsupported servers for ICPM the value to configure as "Maximum Possible Power Consumption (Watts)" from a quickspec PDF?

Thanks in Advance,

Hope it helps.

RCB

[RCB_2]

I forget to say something about, when you setup the "maximum" value, error might still appearing for a while. You can inmediatly override it if you go to Power/Thermal tab and select "Refresh Data" button.

After that, the message is changed and state's something like:

Calculating power history for system %YOUR_SYSTEM%: System %YOUR_SYSTEM% is not powered by a power delivery device that provides power consumption history. The system's history is currently represented by its maximum power value.

This implies that a static value will be considered for calculation purposes.

I still need some help on calculating accurately that value (Mostly for older HP servers as well as other no IPMI Compliant servers from others)

[J_N_Rhodes]

One can also choose to disable a particular alert for all monitored systems

If my post was useful, clik on "White Star" to award me kudos :)

[RCB_2]

How can you achieve that for this particular events? Automated event handling? Can you provide further details about the process?

[RCB_2]

I found also another way to request a "refresh data" for a group of servers instead of individually.

When you define a Datacenter or a Rack, if you ask to refresh data for it, it will do it for every item inside. This is a good approach when defining a lot of servers at a time, as you don´t need to care about them individually.

So, the best recipe (from my point of view after the research) will be:

- Define your datacenters

- Define your racks

- Assign server to the racks

- Review which of the servers have the "Aggregation Error" and check compatibility in the matrix. e.g.

Spoiler

HP Insight Management 7.1 Update 2 Support  Matrix is:  http://h20564.www2.hp.com/portal/site/hpsc/public/kb/docDisplay/?docId=emr_na-c03517804

Go to appendix A:

A Hardware requirements and supported capabilities for

Insight Control power management

and review the "Server Model" in the table.

- If server is supported, troubleshoot it until it's able to aggregate data (Review ILO relation with server, "Enable IPMI/DCMI over LAN" is selected, firewalls etc.)

- If server is NOT supported, figure out which value is better to use as "Maximum Possible Power Consumption (Watts)"

Then go to: TOOLS\Integrated Consoles\Power Management...

There you will have hierarchy defined on prior steps:

Datacenter

                   RACK

                            Servers

You can easily view there the aggregation and communication errors, as well as define the "Maximum" value, faster than looking for each server and tune.

When you are finished defining those values, go to the Datacenter level or Rack level and click on the link in it's name.

Select "Power/Thermal" tab

You can view there the graphic for the rack or Datacenter (which is composed of every reporting item on it)

When you select "Refresh Data", it will ask for a full refresh for the child items.

Afterwards, Clear the aggregation error messages. They will not appear again as error's (Just they will repport aggregation error (with severity:warning instead of severity:error)  because it's using a fixed value instead a "retrieved real" value)

Hope it helps!!! Please, feedback if you found some trouble on the process.

[bolekomp]

Lapinou, anything fix your errors?

 I have the same errors and nothing helps

[shocko]

Are you actively using ICPM?