This section will assist you in performing the routine tasks that will prevent messaging errors, fine-tune performance, and recover lost data when the worst happens.
During normal Exchange operation, there will be times when recipients created on one Exchange server need to be transferred to another. Some common circumstances for relocating recipients include:
Distribution lists and custom recipients are server-independent objects. They belong to a site as a whole, not to a specific Exchange server and therefore, they do not need to be transferred among servers in a site.
Mailboxes are the most common types of recipients. Moving a mailbox also transfers its entire private information store contents (including attachments) to the destination server.
Because the Exchange information store uses intelligent storage techniques for handling attachments addressed to multiple recipients on the same server, additional space is required to split these files. Consider the following example.
User A and User B on server GARLAND01 both receive a message with a 1MB file attachment. The total storage space used on GARLAND01 to hold this file is 1MB. Both users are referencing the same instance of the file.
Suppose you transfer User A to server GARLAND05. The user's entire mailbox and its message contents and attachments are moved to GARLAND05's private information store. This means the 1MB file now exists on both servers.
To move mailboxes to another server in the same site, follow these steps:
FIG. 25.1 Select the destination server in the Move Mailbox dialog box.
To transfer a mailbox to a different site or recipient container, you must use the Import/Export utility provided in the Exchange Administrator program's Tools menu.
NOTE: There is currently no automated method for moving users from one site or recipient container to another. The user's mail could be moved to a personal folder on the user's local desktop or file server home directory while a new mail account is created in the new site or recipient container. The mail could then be placed back on the server in the new mailbox. If you want to use the same mailbox address, follow the steps just given for saving the user's mail, and then export the user using the Directory Export process. Then delete the mailbox in the existing site or recipient container and import the user to the new location. You will need to manually edit the exported data file to correctly import the user into the appropriate container and to allow auto-creation of the new e-mail addresses associated with the new Exchange site.
Normally you will accommodate a wide number of public folder users by creating replicas of folders across your organization. Then, according to the replication schedule, the folders will update each other with changes.
To create a public folder replica, use the following instructions:
NOTE: Select Hidden Recipients from the View menu if your public folders are hidden in the Administrator program display window.
For more information, see Chapter 15, "Information Store Configuration."
Chapter 16, "Creating and Configuring Recipients," goes into more detail about configuring all aspects of the public folder recipient type. The following steps show you how to make changes to the Replicas properties sheet for a public folder.
FIG. 25.2 Select the destination server for this public folder.
CAUTION: Do not replicate the Offline Address Book public folder to any other server in your site. There must be only one instance of that folder in each site. Use caution when replicating any Address Book views that have been customized as well.
The Exchange or Outlook Client (on Windows or Windows NT) can be used to move public folders. If you decide that for browsing purposes or restructuring needs the current hierarchy of Public Folders needs to be redesigned, you can do that only from a capable client.
You must have sufficient permissions on a specific folder in order to move it to a new location. Assuming you have the necessary permissions, move a public folder with the client software by following these steps:
Moving a public folder via the client in this manner is not the same as replicating. The contents of the folder are actually physically moved. If the target folder is on a server in a remote site, the contents will be copied via the client PC to the remote server. This could be very time-intensive, depending on the amount of content in the folder being moved.
NOTE: It is not possible to move Public Folders using the Exchange Administrator program. All moves of this manner can be made only from a capable Exchange client.
A principle part of Exchange maintenance will be adding and deleting mailbox users. When adding a large group of recipients, you can avoid having to enter similar information (such as department or address) by using a mailbox template. You create a template mailbox by creating a dummy mailbox account (one not intended to receive messages) whose detail content you can copy when creating new users.
To create a mailbox template, do the following:
TIP: Check the Hide from Address Book option on the Advanced tab of the mailbox template's properties sheet to prevent it from receiving mail. Then you'll need to temporarily make it visible when you want to use it to import new users (unless you want all your newly created mailboxes to be hidden as well).
To use your template when creating a new user mailbox, follow this procedure:
By using the Directory Import option from the Tools menu, you can create a large number of mailboxes with basic information using a template just like the mailbox that was created in the previous steps.
TIP: If you need to create or modify many users at once, you can use the Directory Import feature of the Exchange Administrator program to import a list of users you have created or previously exported. To test this technique, export a list of your current users. Then, using the same file format, develop your new user list and import that user list. Pay careful attention to the format of the file and note the attributes of each user.
A tool for creating and modifying Import and Export "headers" is available in the Exchange Resource Kit, which you can download from the Exchange Web site.
The information store is the central storage facility for all Exchange messaging data. The information store is made up of the private information store and the public information store. An information store can consume considerable disk space, depending on the size of your organization and the distribution of users per Exchange server. Of course with the massively increased database capacity of Exchange 5.5, these limits and sizes can be very large.
The following are important things to note when maintaining the information stores:
Exchange 5.0/5.5 offers a few new "shortcuts" over the original Exchange 4.0 in order to allow the user to quickly access these status screens from the left pane of the Administrator program without having to drill down through to the properties sheet as noted above. See the additional information that follows about these options.
The Private Information Store Properties dialog box allows you to view the physical resources used by each mailbox on a particular server. This lets you evaluate hard disk space usage and determine storage limits.
First, select the private information store from a server within your site, as outlined in these steps:
Viewing the Physical Resources Used by the Exchange Mailboxes The Mailbox Resources page allows you to view the physical resources (such as hard disk storage space) used by the Exchange mailboxes on a server. It consists of one main display window that shows the mailboxes on the current server. The window is divided into columns of information. You can customize the columns both by the information shown and by their widths on the screen.
NOTE: A mailbox will be displayed in this window only after the first time a user logs on to it. If the mailbox has never been used, it will not appear on the Mailbox Resources properties sheet.
FIG. 25.3 Select the Private Information Store object from among all the other server objects.
There are two ways to view Mailbox Resources. These steps outline the first method:
FIG. 25.4 Monitoring mailbox resources.
NOTE: The Exchange System Attendant and the Exchange service account are always listed in this window.
Columns define what type of information is displayed in the standard dialog box. Default columns are preconfigured to display the most commonly needed information. Optional columns provide more detailed information that may be useful for troubleshooting errors. The following is a list of default columns:
The following is a list of the optional columns you can display in the dialog box:
The second way to view Mailbox Resources is to view them from the Exchange hierarchy (see Figure 25.5).
To access the Mailbox Resources information from the Exchange hierarchy, follow the instruction below:
FIG. 25.5 Mailbox Resources information view.
The Public Information Store Properties dialog box enables you to view the physical resources (such as hard disk storage space) used by the folders held in an Exchange public information store. It consists of one main display window that shows the public folders on the current server. The window is subdivided into various columns of information. You can customize the columns both by the information shown and by their widths on the screen.
The Schedule+ Free/Busy Information and the Offline Address Book are, in essence, public folders as well and are listed in the Public Folder Resources properties sheet.
Viewing the Public Information Store Properties From the Public Folder Resources page, you can monitor the amount of system resources that are exhausted by the use of public folders on your system.
There are two ways to view the properties of the public information store. The first is to select the public information store from a server within your site as outlined in these steps:
FIG. 25.6 View resources utilized by each public folder on this server.
Columns define what type of information is displayed in the standard dialog box. Default columns are preconfigured to display the most commonly needed information. Optional columns provide more detailed information that may be useful for troubleshooting errors. The following is a list of default columns:
The following is a list of optional columns that can be used to display additional information about server resources.
The second way to view public information store properties is to view them from the Exchange hierarchy (see Figure 25.7).
FIG. 25.7 Public Folder Resources information view.
To access the Public Folder Resources information using this new method, follow these steps:
Exchange automatically addresses most defragmentation and compaction issues, but there are times when a manual execution is desired or required. This involves using the Exchange EDBUTIL utility to defragment and optimize disk space allocation for an information store. It is important that you schedule this procedure at a time when message traffic is low or even nonexistent, because it involves stopping the information store service entirely.
NOTE: Software Spectrum has noted that on Exchange servers with the Internet Mail Service installed handling SMTP mail (inbound or outbound), excessive amounts of "white space" in the PRIV.EDB are created as SMTP messages are processed. The automated database clean-up processes do not seem to adequately reduce this fragmented white space. Thus, you should perform a manual defragmentation periodically, depending on the amount and size of SMTP messages handled. The space recovery can be dramatic! One example decreased the size of a 1GB PRIV.EDB to less than 20MB using the EDBUTIL defrag option.
The following command-line entry will defragment and compact the PRIV.EDB file (Exchange Private Information Store) and automatically create a backup file. The EDBUTIL actually performs its work on the backup file, and once its work is satisfactorily completed, EDBUTIL renames the backup to replace the original. This method minimizes possible further corruption and introduction of errors into the current production file. Note, however, that you will need enough disk space to hold both your current EDB file and the backup. (It is possible to copy the file to an alternate location on the network.)
Here is an example command that includes a backup:
edbutil /d /ispriv
In this command, /ispriv indicates that the work is to be performed on the private information store.
The following steps outline the general procedure for compacting the information store:
EDBUTIL [/d] [/information store type]
NOTE: Compacting a large information store could take a reasonably extended amount of time. So make sure to plan for downtime accordingly.
NOTE: While the information store service is stopped, all message transfer attempts to the store are refused. Also, server mailbox users will not be able to open the store through their client.
Option | Description |
/ds | Directory. |
/ispriv | Private information store. |
/ispub | Public information store. |
/b | Backup. Makes a backup of the original file in the specified location. |
/r | Defragment and repair. Defragments the database as usual, but if errors occur, it attempts to correct the problem or removes the error. The end result is a problem-free database--but essential data may be missing. |
/c | Performs a read-only database consistency check without making a copy. It does not repair or affect the database; it simply creates a report. |
/t filename | Renames the newly compacted database with a specified filename. |
Each Message Transfer Agent (MTA) in a site uses the routing table to calculate delivery paths for each message. Whenever a change is made that affects routing (for example, modifying address spaces for a connector), it will not take effect until the site routing table has been rebuilt.
Normally, the routing table is rebuilt once per day unless the Routing Calculation Schedule has been modified. For more information on changing the Routing Calculation Schedule, see Chapter 3, "Exchange's Integrated Server Components." You can also choose to rebuild the table manually, this way:
The MTA retrieves updated routing information from the routing table every 15 minutes. Therefore, it could take up to 15 minutes for routing updates to be used by the MTA. In some cases, it is advantageous to stop and then restart the MTA server to force a "re-read" of the routing table and thus a rerouting of messages in the MTA work queues.
FIG. 25.8 Use the General Message Transfer Agent Properties page to recalculate MTA routing.
The Queues Property page lists messages awaiting delivery by the MTA. There are two primary windows in this property page. The Queue name window shows which queue you are currently viewing and the Message list box displays the messages in that queue. From this, you can view details about a specific message in the queue, change its priority for delivery, or delete it entirely.
Within the Message list window are three columns that contain information about a particular message:
To configure the Queues property page, perform the following steps:
FIG. 25.9 The message queue. It is blank because all pending messages have been delivered.
NOTE: The Internet Mail connector and Microsoft Mail connector have additional queues that can also be accessed through their respective property pages.
When dealing with MTA queue dialog boxes (or any other Exchange queue box), the messages you see on the list are a "snap shot" of all queued messages at one point in time. To get a more dynamic view or to "catch" a message as it passes through the queue, you must continually press the Refresh button to get the latest updates. In some cases, it might be easier to increase the level of diagnostic logging and monitor the Windows NT Event Logs.
Logging falls mainly in the realm of troubleshooting and is discussed in Chapter 28, "Troubleshooting Exchange with Diagnostic Tools." MTA logs can perhaps be used for other purposes, such as billing and accounting for number of messages transferred by person or department if desired.
CAUTION: Increasing diagnostic logging levels on very active servers can result in huge log file sizes in a very short amount of time and dramatically reduce disk performance.
The Microsoft Mail Connector has its own temporary message store called the Microsoft Mail Connector post office. This is sometimes referred to as the "Shadow Post Office." The Microsoft Mail Connector maintains its own message queue property page from which to monitor message queue status.
To open the Microsoft Mail Connector message queue window, do the following:
FIG. 25.10 View Connections.
FIG.25.11 Monitor messages in Queued Messages box.
NOTE: As in the MTA queues property page, the view of queued messages is static. To update the display you must click the Refresh button.
There are two components to each directory object saved in a server's information store: the object itself and a corresponding entry in the directory. Consistency adjustment corrects errors arising by mismatched directory information. This feature will either add or delete a directory entry to match the existence or absence of information store data.
The property page allows you to control at what point these inconsistencies are to be corrected. Figure 25.12 shows the Advanced tab of the Server Properties page where you configure directory inconsistency adjustment.
FIG. 25.12 This property page lets you adjust directory inconsistencies.
Select either All inconsistencies to correct them immediately or select Inconsistencies more than X number of days and enter the time an inconsistency can exist before it is automatically corrected.
Knowledge consistency for the local server is its awareness of other Exchange servers within a site and at other sites within your organization. The knowledge consistency operation is run automatically approximately every 15 minutes. If Exchange servers or sites were added and the local server's directory was not aware of those changes (for example, the local server was down during that time), then you may not want to wait for the automatic checking and do it manually.
Suppose Exchange server CHICAGO01 is brought down for two hours for a memory upgrade. During the time of the upgrade, a new server, CHICAGO08, is added to the site. When CHICAGO01 is restored to proper functioning, it will not be aware of the existence of the new server until the full knowledge consistency cycle is run at the end of the day. Knowing this, the administrator for CHICAGO01 runs the knowledge consistency cycle manually from the Directory Service property pages to make CHICAGO01 aware of the new server's existence. (See Chapter 3, "Exchange's Integrated Server Components," for additional detail on server properties.)
If a directory inconsistency is detected when checking manually, it is a good idea to manually execute all processes related to correcting inconsistencies. Figure 25.13 shows the property page for the Directory Service. By using the following sequence, you can correct knowledge inconsistencies:
Although this sequence is optional and will be automatically executed within a 24-hour period, performing the tasks immediately will reduce the possibility of conflicts or errors.
FIG. 25.13 This property page lets you adjust knowledge inconsistencies
Much like the information store, the Exchange directory can benefit greatly from offline defragmentation and compacting. The Exchange directory service continually performs online compacting during its normal operation, but a dedicated, periodic, offline compacting will provide a more significant increase in performance. Not only will directory access speed improve after compacting, but also hard disk space wasted through inefficient storage will be freed for use. Offline directory defragmentation and compacting requires stopping the Exchange Directory and Directory Synchronization services.
The following command-line entry will compact the DIR.EDB file (Exchange Directory database) and automatically back up the original to a file named DIR.BAK. As with the aforementioned use of EDBUITL with the PRIV.EDB and PUB.EDB, there must be sufficient disk space available to make a complete copy of the current DIR.EDB. Here is an example command:
EDBUTIL /d /ds
where /ds notes that the process is to run against the directory.
The procedure for compacting a directory is as follows:
EDBUTIL [/d] [/information store type] where /d indicates that switch options are present. Please refer to Table 25.1 for the appropriate switch options.
NOTE: Compacting a large directory file could take a large amount of time, so make sure to plan accordingly. For a thousand users, for example, this process will possibly take an hour.
Part of maintaining a healthy Exchange Organization is working with the Windows NT Event Log. It is important to keep track of Event Log sizes (to set upper size limits and expiration dates for log entries) so as to not let the Event Log become too large, but it should be allowed to grow large enough to hold many days worth of log entries. The two Directory Services to watch are these:
Logging levels for each service are set in the Diagnostic Logging property page. See Chapter 28, "Troubleshooting Exchange with Diagnostic Tools," for tips on using the Diagnostic Logging property pages for each service.
Probably the most important maintenance procedure for an Exchange organization is backup of information store and directory information. Few things on a user's computer are as important to him or her as their email. Implementing a good backup strategy is well worth the effort. This section will provide a general conceptual guide to Exchange backup. For more detail in using the Exchange Backup utility, refer to your Exchange server documentation or the documentation of your Exchange "aware" Windows NT backup solution.
Windows NT provides built-in tape backup and restore utilities that are supplemented by extensions installed with Exchange server. These utilities provide the basics for an adequate Exchange backup, but a full-featured backup software package would provide additional functionality and versatility. Note that the Windows NT Backup program has remained essentially unchanged in ability and features from early Windows NT 3.5 days, so usage of a newer tool is strongly recommended.
This section describes backup and restore procedures directly relevant to maintaining Exchange information. For full Windows NT server backups, refer to your Windows NT or third-party backup software documentation.
The version of Backup included with Exchange is aware of the servers in your Exchange Organization and allows you to do backups for a single server, server group, sites, or even an entire organization if necessary.
The main Exchange information you will want to back up is located in the following paths:
\EXCHSRVR\MDBDATA
and
\EXCHSRVR\DSADATA
Important Exchange files to be backed up are as follows:
When restoring lost data, the log files are played back to reconstitute the databases to the point of the last backup.
A proper Exchange aware backup sequence can be completed without any downtime of the server, but performance is affected for connected users. As backup is performed, the transaction logs associated with the main information databases of Exchange (DIR, PRIV, and PUB) are reset to reflect that a backup has been made.
If a "non-Exchange aware" backup tool is used, the Exchange services will need to be stopped in order to release file locks on open files and databases. Once backup is complete, the transaction log files will need to be manually reset.
NOTE: Microsoft has provided a very detailed "White Paper" on backup and disaster recovery giving detailed steps to recover a server or an individual's mailbox from backup. Search the Exchange Web site for a recent version.