You are browsing the archive for Agents.

Avatar of alkin

by alkin

The System Center Management service terminated with service-specific error %%-2147467259

5:19 am in Uncategorized by alkin

I had the following issue on one of my servers

“The System Center Management service terminated with service-specific error %%-2147467259” when I wanted to start my Healthservice.

image

 

When searching the internet I found the following blogpost http://blogs.technet.com/b/smsandmom/archive/2008/04/30/opsmgr-2007-healthservice-service-fails-to-start-with-25362-warning.aspx  . The blogposts says that the State directory registry key can be corrupt but that was fine in my case.

But at the end of the blogpost I found my solution:

This error can be caused by the WindowsAccountLockDownSD Key in at HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\HealthService\Parameters\Management Group\<Management Group Name Here> being invalid or non-present.  And indeed, that was my issue: the WindowsAccountLockDownSD Key was missing on my server.

 

image

The easiest way to resolve the issue with the Windows AccountLockDownSD key is to export the registry key from a similar, working system and then import it in to the registry of the server experiencing the problem.  Once this is complete the HealthService should start successfully.

image

 

thanks,

Alexandre Verkinderen

Avatar of alkin

by alkin

System Center Advisor Changing the upload frequency

11:36 am in Uncategorized by alkin

Most of the configuration like changing upload frequency is done in the registry.

The registry keys for the gateway are stored in HKEY_LOCAL_MACHINE\Software\Microsoft\SystemCenterAdvisor\Gateway and the registry keys for the agent are stored in HKLM\Software\Microsoft\SystemCenterAdvisor\Agent. If you change anything in the registry keys you will have to restart the System Center Management services on the agent and the System Center Advisor Gateway services on the gateway server.

image

By default the gateway only uploads once a day to the Advisor cloud services. Now for demo purposes I’m going to change it to every 12hours.

image

Have a look at the following site http://onlinehelp.microsoft.com/en-us/advisor/ff962521.aspx that describes all the registry keys.

 

Thanks,

Alexandre Verkinderen

Avatar of alkin

by alkin

Fix duplicate relationships for agents to server in Ops DB

9:34 am in Uncategorized by alkin

Sometimes it can happen that agents are ending up with multiple primary management server relationships. Of course you can only have one primary server relationship! In the rare occasions that you end up with multiple primary relationships you can now repair the issue by running the new “Fix duplicate relationships for agents to server in Ops DB” task manually

image

or there is a recovery on this monitor (disabled by default) that you can turn on so that the issue will be fixed automatically:

image

image

 

You will need the latest OpsMgr 2007 R2 management pack that you can find here http://www.microsoft.com/downloads/details.aspx?FamilyID=61365290-3c38-4004-b717-e90bb0f6c148&displaylang=en

Thanks,
Alexandre Verkinderen

Avatar of alkin

by alkin

New KB for OpsMgr Management servers or assigned agents unexpectedly appear as unavailable in the Operations Manager console in Windows Server 2003 or Windows Server 2008

7:25 pm in Uncategorized by alkin

Microsoft just released a new hotfix to deal with ghost agents or agent that are grayed out for an unknown reason. Sometimes it can happen that an agent unexpectedly appear as unavailable without any real reason. We call this ghost agents

The workaround was to delete the health service state directory on the computer (root management server, management server, or agent) that is affected. That way the database will be rebuild.

But Microsoft has now identified this problem as this issue occurs because the database that is used by the health state is corrupted. This database corruption is caused by an issue in the storage engine of the jet database that is hosted on Windows. This is a problem in the Microsoft products and not in OpsMgr. So you need to apply the hotfix to fix the storage engine.

You can download the KB here http://support.microsoft.com/kb/981263 

Notice that this hotfix applies to Windows Server 2008 but not Windows Server 2008 R2.

Hope this helps,

Alexandre Verkinderen

Avatar of alkin

by alkin

OpsMgr : Certificate for this system is not valid when installing Linux agent

3:15 pm in Uncategorized by alkin

Today I ran into some Linux agent deployment issues. I needed to monitor about 20 Redhat Machines . In such an environment environment, Kerberos authentication is not possible. Therefore, certificates are used between the management server and the UNIX-based or Linux-based computers.

windows-to-linux

First if you have some Cross-platform agent deployment issues please have a look at the following blog posts:

 

Ok, let’s start!

So after making sure I had all the pre-requisites needed to deploy an Linux agent I launched the discovery wizard

image

But my agent installation failed because the certificate could not be signed.

image

The certificate signing process does the following:

Operations Manager retrieves the certificate from the agent, signs the certificate, deploys the certificate back to the agent, and then restarts the agent.

image

For an unknow reason my certifcate was not signed and trusted.

 I also got the following error in my event log:

Unexpected ScxCertLibException: Unable to open root store
; input data is: —–BEGIN CERTIFICATE—–
MIIDHjCCAgYCAQEwDQYJKoZIhvcNAQEFBQAwZjEYMBYGA1UEAxMPU0NYLUNlcnRp
ZmljYXRlMTAwLgYDVQQMEydTQ1g2MzMzNzZEMi1FM0UyLTRmMzEtODQ2MS1EMDky

image

 

To solve this problem you need to sign the certificate on your OpsMgr server following this procedure:

 Download and install Winscp on your OpsMgr server.

 Start Winscp and connect to your Linux machine

image

Click yes

image

Browse to /etc/opt/microsoft/scx/ssl

image

Copy the key scx-host-<hostname>.pem  to your opsmgr server.

image

Open the command prompt on your OpsMgr server and change directories to the location where you copied the certificate. Type the command

“scxcertconfig -sign scx-host-<hostname>.pem scx_new.pem”

and then press ENTER. This command will self-sign your certificate (scx-host-<hostname>.pem) and then save the new certificate.

image

Rename your scx_new.pem file with scx-host-<hostname>.ad.pem and replace the original file on your linux server with this file.

image

Connect to your Linux server with putty

image

and type scxadmin –restart

image

This step is very important! If you don’t restart the scxadmin the discovery wizard will still complain about the certificate not being signed!!

 

Now close your discovery wizard and re launch it.

image

The Discovery Wizard discovers the computer and tests to see that the certificate is valid. If the Discovery Wizard verifies that the computer can be discovered and that the certificate is valid, the Discovery Wizard adds the newly discovered computer to the Operations Manager database.Almost immediately you will get a message saying the agent is successfully signed and installed:

image

 

Hope this helps,

Alexandre Verkinderen

Avatar of alkin

by alkin

OpsMgr 2007 R2 MP version 6.1.7553.0 is Released

6:31 am in Uncategorized by alkin

Just got an email from Microsoft saying that they just released their first out of band update to the Operations Manager Management Pack! Notice that this update mp is only available for OpsMgr 2007 R2 environment at the moment, for OpsMgr environment prior to R2 you will need to wait a couple of weeks.

Basicaly they improved the monitoring around excessive agent memory utilization based on customer and blogger feedback they have. 

You can read the full story here http://blogs.technet.com/momteam/archive/2009/10/07/opsmgr-2007-r2-mp-version-6-1-7553-0-is-released.aspx

 

Have fun,

Alexandre Verkinderen

Avatar of alkin

by alkin

Opsmgr Performance problems with healthservice on agents

8:59 pm in Uncategorized by alkin

On some occasions the scom agent can cause a lot of performance issues on the servers. Here I will describe some steps I take when I have performance problems and in most of the cases it works for me.

The first thing you should do is to configure the antivirus exclusions for opsmgr. Have a look at Kevin Holman’s blogpost Antivirus Exclusions for MOM and OpsMgr on how to do this. Configuring anti virus exlusions will help you gain some performance.

This is a screenshot of the performance of my cpu before the anti virus exclusions:

clip_image002[1]

 

And this one is after the anti virus exclusions:

clip_image002[3]

 

The second thing is to have a look at the healthservice.exe. We’ve created a collection performance rule in the previous post to collect the processor time % of the healthservice. If the healthservice is consuming more than 15 a 20 % of your cpu you have a problem . In my case I  had huge performance problems due to the healthservice.exe that was taking all the processor time. My healthservices.exe was taking about 50 to 60 % with peaks to 100% on my agents….You can imagine the reaction of my client….he was not very happy.

Performance view of my healthservice.exe on one of my agents:

clip_image001

So I did some tests, installed all the necessary hotfixes, deleted the healthservice cache, exception of programfiles\scom\ for the antivirus etc etc. But with no results.

So I tought it was WMI and did the following steps to rebuild the wmi:

· Net Stop WinMgmt

· Ren %WinDir%\System32\Wbem\Repository %WinDir%\System32\Wbem\OldRepository

· Net Start WinMgmt

Still no success, still had performance troubles. I contacted one of our company architects who send me a repairwmi.cmd “program” he developed . It will recompile all the mofiles, re-register all the dll’s, exe’s etc.

Steps to take:

1. Copy the content of the repairwmi.cmd to the %WinDir%\System32\Wbem\ folder

2. Run the repairwmi.cmd

3. Stop the opsmgr healthservice

4. Delete the healthservice cache

5. Start the opsmgr healthservice

And voila! All my performance problems where gone!!

clip_image002

clip_image003

clip_image004

clip_image005

clip_image006

 

As you can see this reduced drastically the processor time % of the healthservice!

 

Hope this helps,

Alexandre Verkinderen

Visit Us On TwitterVisit Us On Linkedin