Microsoft has decided to replace the buggy version of the Monitoring Management pack for Exchange 2010 with a new release of the software. Among the errors fixed in the reprise of the service pack (version 14.03.0038.004):
- Occasional double logging of alert events;
- Incorrect reading of values;
- Logging events on local server where Terminal Services is running; and
- Erroneous mass quarantine of mailboxes.
The quarantine error was discovered in June following the release of Monitor Management for Exchange 2010 Service Pack 2. That service pack was designed to address timing problems in a troubleshooting monitor in the package. By fixing that problem, the Exchange team found, it created another.
Once the timing problem was corrected, it exposed a flaw in a storage performance counter. The counter calculates the rate at which logs are being created for a mailbox database. That information is used to assess when a disk will run out of free space. At a specified threshold—the default is 12 hours—mailboxes will begin to be quarantined—as many as 150 at a time—to prevent the mailbox database from dismounting from the disk for lack of disk space.
Such safeguards are good things to have. However, the buggy troubleshooter’s free space calculations in the Monitor Management software were unreliable, causing quarantine features to kick in prematurely.
“This release should not cause the mass-quarantine issue,” Brian Day, a program manager in the Exchange customer experience group, commented in the The Exchange Team Blog.
The first Monitoring Management pack to support Exchange Server 2010 SP1 was released in April 2011. It contained rules and scripts to monitor and report on the performance, availability and reliability of all Exchange server roles—maiboxes, client access, hub transport, edge transport and unified messaging.
The software monitors Exchange events and provides a central location for collecting component-specific performance counters. It uses that information to alert administrators when it encounters events that require operator intervention. Those alerts for service outages, performance degradation and health monitoring can prevent possible service outages or configuration problems. They allow an administrator to proactively manage Exchange servers and identify issues before they become critical
Improvements included in the latest version of the Monitoring Management pack include:
- Resolved ObjectNotFoundExceptions in correlation engine. Prior version of the Correlation Engine could encounter ObjectNotFoundExceptions on a regular basis. The number of exceptions of this type is significantly reduced in this update.
- Reduced load on RMS/MS. A number of improvements were made to reduce the load of the Management Pack on the RMS/MS. They include reducing the the number of read operations the Correlation Engine makes to the SDK to get entity and monitor states, improving the engine’s cache handling when Management Pack updates are applied and increasing the correlation interval time from 1.5 minutes to five minutes.
- Reduced load due to discovery. The discovery interval was increased from four hours to 24 hours and handling of domain controller objects was improved to decrease churn.
- Improved Database Copy Health monitoring. A single Healthy Copy monitor is used to decrease load on RMS.
- Improved performance monitoring. Non reporting Perf Instances are now enabled by default and some write operations were removed to decrease unnecessary writes to the database.
Microsoft Tries Again with Exchange Server 2010 Monitoring Management Pack
Free ebook download: Top 10 Most Popular Troubleshooting Posts for Email Administrators