Error -1206 JET_errDatabaseCorrupted: a non-database file or a corrupt database

Fix it now

Microsoft publishes -1206 as JET_errDatabaseCorrupted: there is a non-database file or a corrupt database. This is the engine reporting on the database file itself, so the first moves are to establish the state of every copy and get a clean one back, not to run mailbox repairs.

Get-MailboxDatabase -Status | Format-List Name,Mounted,EdbFilePath,LogFolderPath
Get-MailboxDatabaseCopyStatus * | Format-Table Name,Status,CopyQueueLength,ReplayQueueLength,ContentIndexState

1. Establish scope from the entry that reported it. A database-level engine error is a database problem; do not read it as damage to one mailbox.
2. If a healthy copy exists in a DAG, activate it. That returns every mailbox immediately and turns the damaged copy into something you can reseed on your own schedule.
3. With no healthy copy, restore from the last known good backup and replay the logs. That is the route that returns all of the data.
4. Confirm the file the engine is complaining about is the file you think it is, and that the paths on disk match what Get-MailboxDatabase reports.
5. Only when there is no copy and no usable backup does a repair enter the conversation, and then against a copy of the file, never the original.

Mailbox repair requests do not address this code. They target logical corruption inside a mailbox, which is a separate problem with its own symptoms, and running them here costs time you do not have.

If a healthy copy is mounted and users are working, you are done for now. The next section explains what -1206 is and how mailbox-level corruption differs.

Why it happens

Microsoft’s published description for -1206 is short and worth taking literally: JET_errDatabaseCorrupted, there is a non-database file or a corrupt database. It is an engine-level condition raised about the file, normally surfaced by the database utility or by a mount that fails. It is not a report about one mailbox, one folder or one item count, and there is nothing in the code that says the rest of the database is fine.

That distinction changes the whole response. For a database-file condition, the reliable moves are the ones that get you a different copy of the data: activate a healthy copy, restore from backup, or as a last resort extract what can be extracted from a repaired copy of the file. Mailbox-level logical corruption is the other thing, with its own signature – wrong item counts, folders that will not open, search folders pointing at nothing – and its own tool in New-MailboxRepairRequest.

The two codes that often travel with -1206 are also published, and they describe record-level complaints rather than file-level ones. -1601 is JET_errRecordNotFound, the key was not found. -1605 is JET_errKeyDuplicate, there is an illegal duplicate key. Seeing those alongside a failed mount tells you the engine is unhappy about structures inside the file; it does not turn a corrupt database into a per-mailbox problem.

Full reference

The codes, as Microsoft publishes them

The last row is there for contrast. A checksum failure is a specific page failing verification on read, and it has its own article. -1206 is a broader statement about the file, which is why the recovery routes are broader too.

Mailbox repair requests: what they are for

New-MailboxRepairRequest detects and fixes corruption inside a mailbox or across all mailboxes in a database. The corruption types you can ask for include SearchFolder, ProvisionedFolder, MessageId, AggregateCounts and FolderView among many others, and Microsoft marks a large number of them – AggregateCounts and FolderView included – as values that can only be used by themselves. That is why the four-value command that circulates in older material fails: it was written against a version whose rules were different.

New-MailboxRepairRequest -Mailbox user@example.com -CorruptionType SearchFolder,ProvisionedFolder -DetectOnly
New-MailboxRepairRequest -Mailbox user@example.com -CorruptionType AggregateCounts
New-MailboxRepairRequest -Mailbox user@example.com -CorruptionType FolderView

Microsoft’s own statement is that after you begin a repair request, you cannot stop it unless you dismount the database. Plan the window on that basis. Remove-MailboxRepairRequest removes request entries; it is not documented as a way to abort a repair that is already running.

When a move is better than a repair

Where a database is mountable but untrustworthy, creating a new database and moving mailboxes into it with New-MoveRequest rebuilds every structure in the target and leaves the old file behind. You can verify each mailbox as it lands and read the move report for items that were skipped. It takes longer in wall clock time than a repair and considerably less of your nerve, and it is the route that leaves you with a database you are willing to keep.

A note on the engine utility

Microsoft no longer publishes a current reference page for the Exchange database utility, so its switches are not presented here as documented syntax. If you reach the point of inspecting a header or attempting a repair, run the tool without arguments to see the mode list for your build, work against a copy of the file, and treat a hard repair as an extraction step: whatever comes out of it must be defragmented, emptied into a freshly created database, and the repaired file discarded.

When it keeps coming back

• Look at the storage. A file-level condition that recurs on the same volume is a storage conversation, not a database one.
• Apply Microsoft’s published antivirus exclusions on every server, active and passive, and check with fltmc filters that nothing else is in the path.
• Check the build. Exchange Server 2016 and 2019 went out of support on 14 October 2025, so a defect fixed in a current build may have no fix available on yours.
• Move the mailboxes to a different database and see whether the problem follows them. If it does not, the file was the problem.

Every code this article covers

Confirm the fix worked

1. A healthy copy is mounted, or the restored database mounts and stays mounted.
2. Get-MailboxDatabaseCopyStatus shows every copy healthy with queues near zero and a healthy content index state.
3. Users can open mailboxes on the database and send through it.
4. If you ran a detection pass on a mailbox, the report shows what was found before anything was changed.

Questions people ask about this