Exchange 2010: Partition/disk of the database is full, what to do?

What can I do if the partition/hard disk on which the Exchange database is stored is full and the information store refuses to work due to insufficient free space?

Experience has shown that it is not the size of the Exchange databases that fills up a partition, but the log files of the databases take up the free memory over time.

But why?

Exchange saves an entry for each action in the database in a transaction log of the database, each transaction log is 1 MB in size, when a log is full, a new one is started.

Why does Exchange do this?

Should the database ever have a defect or inconsistency, the transaction logs, which contain all changes to the database, can be used to restore the database to a consistent state (see also here: https://www.frankysweb.de/?p=342). Exchange therefore saves logs until...

How do I get rid of the log files?

... until the database is backed up. As soon as the database has been backed up, there is no longer any reason to keep the log files. Exchange then deletes the log files. However, this _ONLY_ works if the method or program used to back up Exchange is also Exchange compatible. Software that creates a snapshot of the Exchange VM or an image of the Exchange server must also inform Exchange that the logs can be deleted. If the backup software does not receive this information, logs are written until even the largest hard disk is full.

Example:

In this picture you can see a database (green) with a size of 72 GB, the database has not been backed up for 2 days, therefore there are also many log files (orange) for the database. In total, there are already almost 1900 log files with a size of 1 MB each within 2 days.

Within 2 days, almost 1.9 GB of log files have already been written. Incidentally, the database only contains around 50 user mailboxes. Storage space can therefore quickly become scarce, especially if there are many more mailboxes within the database or the backup does not run for 2 weeks.

But what can be done when the volume is full?

Option 1:

The safest way to get rid of the logs is to back up the Exchange Server with a compatible backup program. The Windows Server Backup can be used quickly for this purpose. This is also the easiest way: Install the Windows Server Backup feature, configure a one-time backup (VSS settings - Full VSS backup), back up to a network drive or USB hard disk, done. The logs have then been deleted and storage space should be free again. The backup is configured as follows:

Select user-defined backup

Then select the partition with "Add elements" that contains the Exchange databases. The following must be selected under "Advanced settings"

In the VSS settings, it is important to select "Complete VSS backup", otherwise the logs will not be deleted. In the last step, the backup target is selected and you are ready to go.

Now you have to wait until the backup is finished, which can of course take some time if there are a lot of logs and a large amount of data. If it has to be done quickly, there is another option.

Option 2:

If the database is in the clean shutdown state, i.e. has been shut down cleanly by Exchange, then all logs have already been imported into the database. It is therefore first necessary to check whether the database has been shut down cleanly. This is done with the command "eseutil" on the command line:

eseutil /mh "NameDerDantenbank.EDB

If the value at State is "Clean Shutdown", everything is OK and you can continue if the value at State corresponds to "Dirty Shutdown", please continue reading here. The next step is to move all transaction logs with the extension .log to another volume or network drive. After moving, there should be space on the volume again. Now the circular logging of the database must be activated so that Exchange does not "miss" the moved log files.

Circulation logging can be configured in the properties of the database by simply ticking the box

The database can then be mounted again. However, it is recommended to use the backup method; the log files should only be deleted in an extreme emergency. Circulation logging should also be switched off again immediately.

And what can I do to prevent this from happening again?

Regular backups of the databases with compatible software reduce the risk of the hard disk filling up with logs considerably. It also makes sense to monitor the volumes for free disk space. If no monitoring solution is used, the "Resource Manager for File Servers" from Server 2008 R2 can also be used to receive a warning when disk space is running low. The Resource Manager does not replace monitoring software, of course, but in this case it is better than nothing.

The "File Server Resource Manager" role service can be added as a role service for the "File Server" role in the Server Manager

The volumes on which the storage space is to be monitored can already be selected during installation

The reports can be configured by clicking on "Options"; all reports can be deselected here. The usage threshold is set to 85 %, i.e. a notification is sent when the volume is 85%iger full. The value that makes sense here depends on the size of the volume. In the next dialog, you can also select that a warning or the report should be sent by e-mail

After installing the role service, the settings can be configured in the "Resource Manager for File Servers" console. First a few general options, such as sender and default recipient, see the following screenshots

You must then tick the box to send a warning by e-mail when the storage space is running low.