OpsMgr 2007 Agent Deployment Footprint

不错的文章，很有参考价值，拿出来share一下.

Purpose:
We decided to test the impact of an installed agent on a managed machine. The following are results from tests performed to determine the impacts on the processor, network, disk, and memory on a client system monitoried by Operations Manager 2007.
Our test environment ran on two different virtual servers running on two different guest operating systems (details regarding the test environment are at the end of this blog entry).

Please note that these tests were performed on low-end server configurations so that the maximum impact on the systems should be able to drawn from this, e.g. a worst-case scenario. Higher-grade systems should be less impacted than these particular systems were.

The installation ran on the two systems for between 3 and 3 1/2 minutes. We stopped gathering statistics when the agent listed as successfully deployed from the Management Server. During installation we noted the following ranges of impacts on the system:

Processor: 29-37% increase during installation
Disk: 18-19 additional pages per second
Disk: 154 MB data (%programfiles%\System Center Operations Manager 2007)
Network: 11 MB data sent to the system during installation
Memory: 22 MB less available memory

After the installation was complete, the performance counters gather prior to the installation were compared to those gathered after installation. The results indicate what should be the worst-case impact of the agent on the system.

Processor: 3-4%
Disk: 0.5 additonal pages per second
Disk: 156 MB data (%programfiles%\System Center Operations Manager 2007)
Network: .5-.6 MB/min additional traffic
Memory: 11 MB less available memory

Testing environment specifications:
To test the impacts of both the deployment itself and the agent functionality after installing the agent, we took a total of three sets of performance counters (from perfmon and netmon). One was prior to installation to provide the baseline, one during installaton and the last after installation was complete.

Host Operating System #1 running on a dual processor 1ghz system with 2GB memory. Host OS running only Virtual Server 2005 R2.
Guest Opeating System #1 running within Virtual Server 2005 R2 with 1GB memory.

Host Operating System #2 running on a single processor 1.5ghz system with 2GB memory. Host operating system running only Virtual PC 2007.
Guest Operating System #2 running within Virtual PC 2007 with 1GB memory.

Both guest OS running Windows Server 2003 without either a domain controller or Exchange installed on the system.
The Operations Manager 2007 environment is running on a Root Management Server separate from the Operations Database Server, and has all currently available management packs loaded.

Two bugs found in SCOM 2007 RTM

测试时发现了2处Bug, 都是关于Web console的。

1. 无法手工输入account在web console里。临时解决方法为runas or change IE settings in intranet zone.
2. 不同account的视图在web console 里会造成mangement server CPU overload 100%. 暂无解决方法.

Monitor a logfile with Ops Mgr 2007

找到一篇关于Monitoring txt log file in SCOM 2007, 比较简单, 稍后会测试一下。

I saw a question about monitoring a logfile with Ops Mgr 2007. Here is a short step by step guide
In the console, under Authoring, Management Pack Objects, Rules, create a new rule
Choose to create a Alert Generating Rules/Event Based/Generic Text Log (Alert), click next
Input a rule name and choose rule target, click next
Input a directory, for example C:\, input a pattern, for example logfile.log, click next
Input the following expression,Parameter Name = Params/Param[1]Operator = Matches wildcardValue = warningclick next
Configure a suitable alert and then click create
Please note that this rule is case sensitiv, so it will make difference between WARNING and warning.

Good News for MOM Connectors !!

Microsoft Acquires Engyro Corporation
CINCINNATI, Ohio - June 4, 2007 - This week, Microsoft has acquired Engyro Corporation, a successful partner company who provides software solutions that help organizations integrate, consolidate and simplify their IT operations, extending the enterprise interoperability of System Center Operations Manager 2007 with other enterprise management systems. This alliance will further Microsoft’s commitment to its customers’ longer term need for interoperable enterprise solutions and further the System Center family of solutions including Operations Manager and System Center Service Manager when it is released.
“We are very pleased with the acquisition of Engyro by Microsoft as it now brings enterprise interoperability to Operation Manager customers. Our technology allows Operations Manager to interoperate with different enterprise management systems providing the ability to share alert/event information across system management tools. The customer gains the value of integrating best of breed management for Windows and Microsoft technologies with their non-Microsoft management tools. We believe that every current and future Engyro customer will benefit from this alliance.” said Jim Dennedy, Engyro President and CEO.
more pleae click here

Resetting Health States in OpsMgr 2007

看到了一篇关于Reset Health State的文章，比较有帮助。

Resetting the Health State on the Root Management Server
The Operations Manager 2007 Root Management Server in an environment we were working on showed up in the console as Yellow, and going through the Health Explorer there were a variety of Red and Yellow items within it. There were issues during the original installation which caused database communication issues, and were thought to be the cause of the issues being seen. No alerts could be seen to be causing this state to be Yellow.
We tried a variety of methods to reset the health state of the RMS, but none of them succeeded (server reboots, Reset Health in Health Explorer, Maintenance Mode, Flush Health Service State and Cache). Thanks to some direction provided by the MVP community (Thank you Gordon!) we found that the following worked as steps for us to reset the health state of our RMS and allowed us to determine if the issues seen were historical or current.

· Launch Health Explorer for the RMS and document any Yellow or Red monitors within the Root Management Server.
· Disable Monitors which are Red using an Override which applies only to the RMS server.
· Put the RMS server into Maintenance Mode.
· Wait until the server is unmonitored.
· Take the RMS server out of Maintenance Mode.
· Re-enable the Monitors which were disabled.
· Launch Health Explorer and refresh the state.

Resetting the Health State on other Agents
You don’t have to do the above process with regular agents. If the Reset Health task within the Health Explorer does not work then for the non-RMS servers you can use the Flush Health Service State task. It’s available under the monitoring \ Operations Manager \ Agent Health State. On the right-pane click on the server, choose Health Service Tasks \ Flush Health Service and State Cache. This isn’t available for the RMS server itself though.