This post covers the nuances of long-term backup of PST Files, their backup using Azure Backup, a component of Microsoft Operations Management Suite (OMS), and ways to check they’re backed-up the right way.
Why use PST files?
Mailbox size management
Mailboxes located on Exchange Server are configured to have limited sizes by most Exchange admins. PST Files provide a mechanism to simply drag and drop all mail assets from the mailbox to a PST File on a local hard disk, ensuring you always have space on your mailbox to send and receive emails.
PST files allow you to store mail assets important to you so you can take them along as you move across mailboxes, for example as you move across organizations or even as a reference to old email.
In enabling mail users to store email data over long periods of time for future reference, these files, whether they are on the exchange server or local hard-disk, require protection from unexpected disk/server failures, making them good candidates for storing in a long term cloud backup solution like Azure Backup.
Challenges with long-term backup of PST files
PST Files are large in size (generally 2GB to 20GB). Whenever Outlook.exe or processes that depend on Outlook.exe are running, the associated PST Files are opened and locked for writes by Outlook. Given that Outlook runs continuously, PST files are mostly open. These factors pose several challenges making robust backup of PSTs a hard problem to solve.
Challenges due to large-size of PST files and reducing backup-interval
- Backups need to be done in a space-efficient manner
- Backups need to be done quickly without consuming high bandwidths
Incremental backup, which backs up only the changed portions of a file or disk serves to be the most optimal way to reduce backup storage and perform backup quickly.
Challenges due to PST files being mostly open
- The backup application cannot access the file as it is locked by the application associated with it, making the backup solution spawn multiple re-tries but eventually failing to perform consistent backups of the PST File
- If the incremental approach is not possible at all and the backup application relies on modified flag of the PST file, it will always look modified at every backup attempt, leading to the entire file being sent again for backup
Microsoft Azure Backup of PST files
For robust and optimized backup of PST files, the backup product should have the following minimum features.
- Track block level changes on the volume corresponding to changes in files
- Take volume snaps such as using Windows Volume Shadow-Copy Service (VSS) APIs
For reliable tracking of block level changes, Microsoft Azure Backup products use either a kernel mode component called File System Filter Driver or maintain checksums for changed blocks on a file. Both approaches use OS documented features and APIs to intercept file I/O without blocking applications associated with the file, to store the information about only the changes on any file (such as PST) in a volume.
Microsoft Azure Backup products also rely on VSS APIs to take a volume snapshot that creates a virtual snapshot of the regular volume that the backup application uses to backup files while associated applications continue to work on the regular volume. The shadow-copy creation request spawns a series of OS-aware steps to ensure I/O consistency of the files on the volume, which is required for open files.
An important point to note here is that the full set of VSS APIs for reliable shadow-copy creation is only possible when the caller is in the administrator mode which is default for Azure Backup products. Some providers require users to manually configure this and selectively enable services that allow VSS APIs to be called in the recommended manner.
Once the volume-snapshot is taken, the backup-application then uses the stored block-level change information to only send the changed portions of the file on the snapshot for backup. This method for backup ensures the following.
- Backups are incremental and therefore space-efficient
- Backups finish quickly without consuming large-bandwidths due to large size of the PST Files
- Backups are I/O consistent and have the latest changes despite the file being locked by Outlook or any other associated application
How to check if your backup product backs-up PST files the right way
Having identified the nuances of robust incremental backup of PST Files, it will be valuable to run a few checks to see if your current backup application can backup PST Files the right way:
Check for UAC prompt during Installation, Launch, Upgrade
Given that the backup product should contain kernel mode components and also run in administrator mode for full access to VSS API, you must encounter a User Account Control (UAC) prompt either during Installation, Launch or Upgrade of the application.
If you did not encounter a UAC Prompt, chances are your backup product does not take volume-snapshots and might not even track changes on PST Files, leading to unreliable backups.
Check the size of backup data transferred
After setting up your application, check the status of backups of PST files. If you aren’t seeing any backup failures during hours when your outlook was open and the size of the data transferred is considerably less than the size of the PST File, then you can be more or less sure about the expected backup of your PST Files.
Several products claim they do incremental backups in general, but fail to mention that for always open files such as PST, they default to backing up the entire file each time. This is generally due to the inability to intercept file I/O reliably to track block-level changes. In such cases, although the backup application knows that the file has been modified, it can’t identify changes made in the file. This leads to the entire file being sent for backup.
Start protecting your PST files
If you are an existing Azure customer and haven’t been protecting your PST Files, install the latest Azure Backup agent on your windows machines now to start protecting them. If you are new to Azure, get started with free Azure trial subscription.