Monitoring Exchange Server 2010 : Monitoring Exchange Databases (part 1) - Monitoring Exchange Database Information and Statistics

Microsoft recommends monitoring your databases on a daily basis. The main advantages to daily monitoring are as follows:

• You can quickly detect and address issues that may affect the messaging service or data availability.

• You can ensure the successful completion of specific administrative tasks, such as daily backup operations.

• It helps you meet the requirements of your Service Level Agreements.

In this lesson, you consider how you monitor public folder and mailbox database statistics to detect anomalies and indications of current and future problems and how you check the status of an Exchange Server 2010 database. You also consider how you monitor status information about mailbox database copies included in a database availability group (DAG).

1. Monitoring Exchange Database Information and Statistics

In order to check the health of your Exchange databases and of the replication process (if configured), you need to monitor your databases. You can obtain general information about both public folder and mailbox databases and statistics related to the public folders and mailboxes they contain. You can obtain more specific information about mailboxes and mailbox databases, such as usage information, information about queues, information about lagged mailbox database copies, and information about disconnected mailboxes.

1.1. Obtaining Information about Mailbox Databases

When you are monitoring a mailbox database you first need to obtain general information about the database, such as the server where it is located, its mailbox retention period, deleted item retention period, quota limits, associated public folder database, and so on. You can use the Get-MailboxDatabase cmdlet to obtain general information for mailbox databases. For example, the following command returns detailed information about all the mailbox databases in an Exchange 2010 organization:

Get-MailboxDatabase | FL

The previous command typically returns a lot of information that you would likely redirect into a text file for analysis. In a large organization that has a lot of databases, this command returns an excessive volume of information. You have the option of refining this information by specifying the Mailbox server, the mailbox database identity, or both. For example, the following command returns detailed information about the mailbox database named Mailbox Database 1363123687:

Get-MailboxDatabase -Identity "Mailbox Database 1363123687" | FL

Figure 1 shows some of the output from this command.

You can obtain statistics (if available) about the transport dumpster, such as dumpster deletes per second, dumpster inserts per second, and dumpster item count by including the DumpsterStatistics parameter, as demonstrated by the following command:

Get-MailboxDatabase -Identity "Mailbox Database 1363123687" -DumpsterStatistics | FL

					  

You can also determine the status of the mailbox database, if this is available, by including the Status parameter. This tells you whether the status of the mailbox database is one of the following: BackupInProgress, Mounted, or OnlineMaintenanceInProgress. It also tells you the available free space in the database root. Checking the status of mailbox databases is an important step in monitoring replication health. The following command uses the Status parameter:

Get-MailboxDatabase -Identity "Mailbox Database 1363123687" -Status | FL

Note:

THE DATABASE ROOT

The Extensible Storage Engine (ESE) organizes database storage in a three-level hierarchy: database root, tables, and indexes and long values. The database root owns all the space in the database. Tables request chunks of space, which they then own in conjunction with the database root. Index and long-value trees request space from a table that in turn owns space allocated from the database root.

You can also use the EMC to determine the properties of a mailbox database. The procedure is as follows:

The tabs on a mailbox database Properties dialog box are General, Maintenance, Limits, and Client Settings, as shown in Figure 2.

1.2. Viewing Mailbox Statistics

You can use the Get-MailboxStatistics EMS cmdlet to view the statistics for all the mailboxes on a server, for all the mailboxes in a mailbox database, or for a single mailbox. Note that you cannot use the cmdlet without arguments to view the statistics for all mailboxes in an Exchange organization. The cmdlet requires the Server, the Database, or the Identity parameter.

For example, the following command lists the statistics for all the mailboxes on the VAN-EX1 Mailbox server:

Get-MailboxStatistics -Server VAN-EX1 | FL

The following command lists the statistics for all the mailboxes in the mailbox database Mailbox Database 1363123687:

Get-MailboxStatistics -Database "Mailbox Database 1363123687" | FL

Figure 3 shows the output from this command.

If a user mailbox has been created but has never been accessed, that mailbox is not included when you list the statistics of mailboxes on a server or in a mailbox database. If you attempt to obtain statistics for a mailbox that has not been accessed, you will get no statistical information but will instead receive a warning message. Suppose, for example, that the Don Hall mailbox has never been accessed and you enter the following command:

Get-MailboxStatistics -Identity "Don Hall" | FL

This results in the message shown in Figure 4.

If a mailbox returns statistics, you can use the PowerShell format-list (FL) cmdlet to display the value of one or more specified statistics. This is more convenient than searching through the list of all the statistical values. For example, the following command displays the last logon time for the Kim Akers mailbox:

Get-MailboxStatistics -Identity "Kim Akers" | FL LastLogonTime

You can use the Sort-Object PowerShell cmdlet to sort the mailboxes in a mailbox database or on a Mailbox server by the value of one or more mailbox statistics. You can do this in either descending or ascending order and use the format-table (FT) PowerShell cmdlet to display the results as a table. For example, the following command lists the mailboxes in the Research mailbox database in descending order of item count:

Get-MailboxStatistics -Database Research | Sort-Object ItemCount -Descending | FT
DisplayName,ItemCount

					  

The following command lists the mailboxes in the Research mailbox database in descending order of total item size:

Get-MailboxStatistics -Database Research | Sort-Object TotalItemSize -Descending | FT
DisplayName,TotalItemSize

					  

Note:

MAILBOX SIZE

The size of a mailbox is the sum of its total item size and total deleted item size. This is not returned directly as a statistical value. If you want to display total mailbox sizes, you need to use a PowerShell routine that combines the TotalItemSize and TotalDeletedItemSize statistics for this mailbox.

If you do not want to list all the mailboxes in a mailbox database or on a Mailbox server but instead want to list, for example, the top five mailboxes in terms of total item size, you can pipe the results of your search into the Select-Object PowerShell cmdlet. For example, the following command lists the top five mailboxes in Mailbox Database 1363123687 in descending order of total item size:

Get-MailboxStatistics -Database "Mailbox Database 1363123687" | Sort-Object
TotalItemSize -Descending | Select-Object -First 5 | FT DisplayName,TotalItemSize

					  

The following command lists the top 10 mailboxes on the Mailbox server VAN-EX1 in descending order of item count:

Get-MailboxStatistics -Server VAN-EX1 | Sort-Object ItemCount -Descending | Select-
Object -First 10 | FT DisplayName,ItemCount

					  

Note:

USING THE SORT-OBJECT AND SELECT-OBJECT POWERSHELL CMDLETS

You can use the Sort-Object and Select-Object PowerShell cmdlets with other EMS cmdlets such as Get-MailboxDatabase if, for example, you want to list the top five mailbox databases on a Mailbox server in terms of database size.

Sometimes pressure on a mailbox database can be the result of certain users consuming a disproportionate amount of resources, such as by sending an excessive number of large attachments to a large number of recipients. To detect this situation, you can use the Get-StoreUsageStatistics EMS cmdlet to generate a report on the 25 accounts that are using the greatest amount of resources within a mailbox database. For example, the following command returns the 25 users with the largest mailboxes in Mailbox Database 1363123687:

Get-StoreUsageStatistics -Database "Mailbox Database 1363123687"

You can also use the Get-StoreUsageStatistics cmdlet to obtain statistics about a specific mailbox, but only if this mailbox account is one of the 25 top resource users. Under these circumstances, the following command would generate a report about the Kim Akers account:

Get-StoreUsageStatistics -Identity "Kim Akers"

You can use the Exchange Server Performance Monitor tool to monitor counters that can indicate whether resources in your Exchange organization are coming under stress. This is the same tool as Performance Monitor (PerfMon) except that the Performance and Logs Alerts snap-in has been prepopulated with a large number of Exchange-related performance counters. You can access Exchange Server Performance Monitor from the EMC by clicking the Toolbox node, clicking Performance Monitor, and clicking Open Tool.

As an experienced administrator, you should have used PerfMon on both server and client operating systems to monitor current performance, create data records, and generate reports. The tool provided by Exchange Server 2010 works in the same way, but you have the option of monitoring performance counters specific to Exchange. More than 100 MSExchange countergroups exist, each one of which offers a significant number of counters.

You are not expected to be familiar with every counter that populates the Exchange Server PerfMon. It is probable that no single person could list and describe all of them. Nor can you be expected to know the acceptable values that every counter returns. Although some countervalues exist that definitely indicate a fault while other values indicate that a process is operating in a satisfactory manner, absolute good or bad values are not typical. If you want to use PerfMon counters to monitor performance, you would normally record baseline counter values when your Exchange organization is working in a satisfactory manner—possibly during a quiet period, under normal conditions, and during busy times. Recording values on a regular basis under the same conditions and comparing these to baseline values would indicate whether performance is deteriorating over time and help you track trends.

For example, to ensure that your mailbox databases continue to operate efficiently, you need to check that they are being defragmented online on an ongoing basis. Online defragmentation is a background task that operates continuously by default. Exchange Server 2010 provides the following performance counters for monitoring the behavior of online database defragmentation:

• MSExchange Database → Online Defrag Average Log Bytes Shows average size of the log records being generated by online defragmentation

• MSExchange Database → Online Defrag Data Moves/Sec Shows the number of times that data is moved from one page to another by the online defragmentation process

• MSExchange Database → Online Defrag Log Records/Sec Shows the number of times per second that data is moved from one page to another by the online defragmentation process

• MSExchange Database → Online Defrag Page Moves/Sec Shows the number of times that data is moved from one page to a new page by the online defragmentation process

• MSExchange Database → Online Defrag Pages Dirtied/Sec Shows the rate at which online defragmentation is modifying clean database pages

• MSExchange Database → Online Defrag Pages Freed/Sec Shows the number of pages per second that are freed from the database by the online defragmentation process

• MSExchange Database → Online Defrag Pages Preread/Sec Shows the rate at which database pages are read in anticipation of future use by online defragmentation

• MSExchange Database → Online Defrag Pages Read/Sec Shows the rate of database read operations being performed by online defragmentation

• MSExchange Database → Online Defrag Pages Re-Dirtied/Sec Shows the rate at which online defragmentation is modifying database pages that already contained modifications

• MSExchange Database → Online Defrag Pages Referenced/Sec Shows the rate at which online defragmentation is touching database pages

Note:

USING EXCHANGE SERVER PERFORMANCE MONITOR TO MONITOR QUEUES

Exchange Server PerfMon can monitor a large number of Exchange-related counters. In particular, you can use it to obtain information about transport queues by monitoring counters in the MSExchange Transport Queues group, such as MSExchange Transport Queues→ Active Mailbox Delivery Queue Length, MSExchange Transport Queues→ Aggregate Delivery Queue Length (All Queues), MSExchange Transport Queues→ Items Queued For Delivery/Sec, MSExchange Transport Queues→ Items Completed Delivery/Sec, and MSExchange Transport Queues→ Poison Queue Length.

1.5. Obtaining Information about Public Folder Databases

When you are monitoring a public folder database, you first need to obtain general information about the database, such as the server where it is located, what public folders it contains, its maximum item size, quota limits, replication schedule, and so on. You can use EMS commands based on the Get-PublicFolderDatabase cmdlet for this purpose. For example, the following command gets detailed information about every public folder database in an organization:

Get-PublicFolderDatabase | FL

Figure 5 shows some of the output from this command.

If you want to obtain detailed information about a specific public folder database, you can specify the Exchange Server 2010 Mailbox server on which it is located. For example, the following command gets detailed information about the public folder database on the VAN-EX1 server.

Get-PublicFolderDatabase -Server VAN-EX1 | FL

It is typically easier to analyze and store this information if the output of the command is redirected to a text file. For example, the following command redirects detailed information about the public folder database on the VAN-EX1 server to the text file PublicFolderDetails .txt in the DatabaseInformation folder on the VAN-EX1 server:

Get-PublicFolderDatabase -Server VAN-EX1 | FL >
C:\DatabaseInformation\PublicFolderDetails.txt

You can use the Status parameter of the Get-PublicFolderDatabase cmdlet to obtain backup and mount status information (if available). Checking the status of public folder databases is an important step in monitoring replication health. The following command gets detailed information about the public folder database on the VAN-EX1 server, including status information:

Get-PublicFolderDatabase -Server VAN-EX1 -Status | FL

The procedure to view public folder database properties is very similar to that for viewing mailbox database properties described earlier in this lesson. The procedure is as follows:

1.6. Viewing Public Folder Statistics

No EMS cmdlet exists that returns the statistics for an entire public folder database. However, you can use the Get-PublicFolderStatistics EMS cmdlet to obtain statistics for each individual public folder within a public folder database. You would, for example, obtain public folder statistics if you had created a new public folder database on another Mailbox server and you wanted to replicate a public folder on your current server to the new public folder database.

You can check replication by ensuring that, for example, the number and size of items in the public folder replica are the same as in the original public folder. You could use the same procedure if you replicate an entire public folder database and want to check that replication is working correctly.

The following command obtains statistics for every public folder in an Exchange 2010 organization:

Get-PublicFolderStatistics | FL

The previous command can return an excessive volume of information, particularly if you have a large number of public folders. As with mailbox databases, you can refine the information. If, for example, you want to obtain statistics for every public folder in the public folder database on the Mailbox server VAN-EX1, you would enter the following command:

Get-PublicFolderStatistics -Server VAN-EX1 | FL

If you want to obtain statistics for the public folder MyPublicFolder on the Mailbox server VAN-EX1, you would enter the following command:

Get-PublicFolderStatistics -Identity \MyPublicFolder -Server VAN-EX1 | FL

Figure 7 shows the output from this command. Note that this command runs only if you have previously created a public folder named MyPublicFolder as a child of the public folder root.

1.7. Detecting Database Errors

Database corruption occurs, for example, when the Exchange Server Database (EDB) file is damaged during improper server shut down, through virus infection, or because of physical damage to the storage media. If an entire Exchange database becomes corrupt, the ESE writes error messages to Event Viewer, such as “Unable to read database header—database may have moved, or data was moved or was missing during recovery” or “Database page read failed verification because of a 1018 error (page checksum mismatch).” You can use tools such as Eseutil and Isinteg to repair a corrupt database or to fix database integrity problems. However, Microsoft recommends restoring a corrupt database from backup because the use of a tool such as Eseutil can lead to data loss.

Sometimes a database is not itself corrupt but holds items such as email messages, which are. Corrupt email messages can cause problems with mail flow and lead to excessive queuing. If you move an Exchange database to another server or copy its contents to a new database, corrupt messages do not move or copy. You can then delete such messages, typically with a non-delivery report (NDR).