Fix it now
The client asked for public folder content and the server could not reach the mailbox that holds it. The name is a leftover from the era when public folders had replicas; modern ones do not. Each folder’s content lives in exactly one public folder mailbox, so if that mailbox is unavailable, unassigned or excluded from serving, there is nothing else to offer.
Get-Mailbox -PublicFolder | Format-List Name,Database,IsExcludedFromServingHierarchy
Get-MailboxDatabase -Status | Format-Table Name,Server,Mounted
Get-Mailbox user@contoso.com | Format-List Name,*PublicFolderMailbox*
Get-PublicFolder \Path\To\Folder | Format-List Name,ContentMailboxName
- Read both public folder mailbox properties on the user, not just one.
DefaultPublicFolderMailboxis what an administrator pinned;EffectivePublicFolderMailboxis what the user is actually being served by. - If the pinned mailbox no longer exists, hand the user back to automatic load balancing with
Set-Mailbox user@contoso.com -DefaultPublicFolderMailbox $null, or pin them at a live one by name. - Confirm the database holding the failing folder’s content mailbox is mounted. One dismounted database takes out every folder whose content lives there and leaves the rest working normally.
- Have the user restart Outlook. The client caches its hierarchy binding, so a repaired server can keep failing at a client that has not reconnected.
IsExcludedFromServingHierarchy is set deliberately during migrations to stop a mailbox serving clients, and it is routinely left behind afterwards. If every mailbox carries it, nobody can open the tree at all.
If the folder opens, you are done. If it still does not, the next section covers how a client reaches a folder and which link in that chain is broken.
Why it happens
Two separate things have to work. First the client needs the hierarchy, the folder tree itself, which it gets from a public folder mailbox nominated to serve it. Every user is bound to one, either explicitly through the default public folder mailbox property on their own mailbox or automatically, because by default users are distributed evenly across the available public folder mailboxes. Second, when the user opens a folder, they are redirected to whichever public folder mailbox holds that folder’s content, which is frequently a different mailbox entirely.
Only one copy of each folder’s content exists. Microsoft is explicit about it: content is stored in the public folder mailbox and is not replicated across multiple public folder mailboxes, and all users access the same mailbox for the same set of content. There is no replica to fall back on, so if the mailbox holding it is on a dismounted database, has been removed, or cannot be reached from the server handling the request, there is nothing else to offer. The error text talks about replicas because the code path is old.
The hierarchy adds a second failure mode with a different shape. Only the primary public folder mailbox holds a writeable copy of the tree; every other public folder mailbox holds a read-only copy that it synchronises from the primary. That synchronisation runs at least once every 24 hours for each mailbox, and every 15 minutes while users are logged on to a secondary mailbox through Outlook or an Exchange Web Services client. A secondary that has fallen behind serves a tree missing a folder created last week, which produces a folder-not-found result rather than a replica error. That difference is a useful way to tell the two apart before you start changing things.
The database holding a public folder mailbox is not mounted
You have this one if Several folders fail at once, all belonging to the same content mailbox, while other folders open normally.
- Confirm mount state with
Get-MailboxDatabase -Status | Format-Table Name,Server,Mounted. - Mount the database and read the application event log for why it was down.
- In a database availability group, check copy health before assuming the mount is the whole story.
- Have a client reopen the folder once the database is back.
The user is pinned to a public folder mailbox that no longer exists
You have this one if One user or a small group fails while everyone else is fine, and their default public folder mailbox names something you do not recognise.
- Read both properties:
Get-Mailbox <user> | Format-List Name,*PublicFolderMailbox*. - Clear the pin with
Set-Mailbox <user> -DefaultPublicFolderMailbox $nullso automatic distribution takes over, or name a live mailbox instead. - Have the user restart Outlook so the binding is picked up rather than served from cache.
- If many users are affected, check whether a public folder mailbox was removed without reassigning the users bound to it.
Every hierarchy-serving mailbox has been excluded
You have this one if Nobody can open the tree at all, and IsExcludedFromServingHierarchy is set on the mailboxes you expected to serve it.
- List the flag across all public folder mailboxes:
Get-Mailbox -PublicFolder | Format-List Name,IsExcludedFromServingHierarchy. - Clear the exclusion on at least one healthy mailbox so it re-enters the automatic distribution.
- Confirm the primary public folder mailbox is present and its database is mounted.
- Test with one client before announcing the fix.
The flag exists so a mailbox can be taken out of service during a migration. It is not an error state, which is exactly why nobody notices it was left on.
The hierarchy on a secondary mailbox is stale
You have this one if A folder created or moved recently is missing for some users and present for others, and a folder-not-found result appears rather than a replica error.
- Run
Update-PublicFolderMailbox -Identity <secondary mailbox> -InvokeSynchronizeragainst the secondary mailbox. - Add
-FullSynconly if the tree has drifted badly rather than by a folder or two. - Re-test with a user whose effective public folder mailbox is that one specifically.
-InvokeSynchronizer can only be used on secondary hierarchy public folder mailboxes. Aimed at the primary it is not the command you want, because the primary is the writeable copy everything else syncs from.
Permissions rather than availability
You have this one if MapiExceptionNotAuthorized or a folder-not-found result, and the folder opens for administrators but not for the user complaining.
- Read the folder permissions:
Get-PublicFolderClientPermission -Identity \Path\To\Folder. - Add the right the user needs with
Add-PublicFolderClientPermission, choosing the smallest role that works. - Check parent folders too. A user who cannot see a parent cannot navigate to the child.
The published definition of the web services folder-not-found response covers both a folder ID that is not valid and a caller without permission on the folder, so do not read it as purely a hierarchy problem.
Full reference
The properties that decide who serves whom
| Property | Where it lives | What it means |
|---|---|---|
DefaultPublicFolderMailbox |
On the user’s mailbox | The public folder mailbox an administrator pinned. Empty means automatic distribution |
EffectivePublicFolderMailbox |
On the user’s mailbox | The mailbox actually serving that user right now |
IsExcludedFromServingHierarchy |
On a public folder mailbox | Takes that mailbox out of the automatic distribution. Users pinned to it explicitly still reach it |
ContentMailboxName |
On a public folder | Which public folder mailbox holds that folder’s content. The only copy there is |
IsRootPublicFolderMailbox |
On a public folder mailbox | Identifies the primary, which holds the one writeable copy of the hierarchy |
Hierarchy synchronisation timing
- Each public folder mailbox has the hierarchy synchroniser and the mailbox assistant invoked at least once every 24 hours.
- While users are logged on to a secondary mailbox through Outlook or an Exchange Web Services client, the synchroniser runs every 15 minutes.
- Write operations against the hierarchy are synchronised immediately, so a folder created by an administrator is not the slow case.
Update-PublicFolderMailbox -InvokeSynchronizeris for troubleshooting; it is not something to schedule.- Without
-SuppressStatus, the command prints status every three seconds for up to a minute and blocks the shell while it does.
Finding the single copy
When a specific folder fails and its neighbours do not, the question is only ever which mailbox holds it. Get-PublicFolder \Path\To\Folder | Format-List Name,ContentMailboxName answers it, and from there you are checking one mailbox and one database rather than the whole estate. If several unrelated folders fail together, list their content mailboxes and look for the one they share.
Adding capacity rather than repairing it
Where the fault is load rather than availability, Microsoft’s documented route is to create another public folder mailbox and move users onto it: New-Mailbox -Name "<name>" -PublicFolder, then Update-PublicFolderMailbox -InvokeSynchronizer -Identity "<name>" so it has the tree, then either pin users at it or clear their pins so automatic distribution includes it. Clearing pins is usually the better move, because a pinned user does not benefit from any mailbox you add later.
When the server side is clean and one client still fails
- Restart Outlook. The hierarchy binding and the folder tree are both cached client-side.
- Check whether that user is pinned to a mailbox while their colleagues are not.
- Confirm the user’s effective public folder mailbox is one that is mounted and not excluded.
- Test the same folder from a second machine signed in as the same user, to separate the profile from the account.
- Check the application event log on the server holding the content mailbox rather than the one the client connected to.
When a licence is the actual fix
Nothing here needs a purchase. Reassigning a public folder mailbox, clearing an exclusion flag, adding another public folder mailbox or mounting a database all cost nothing and work today. The licensing question is a different and larger one: public folders are frequently the last workload keeping an on-premises server alive, and the hierarchy routing described above is the price of keeping it. If the content is mostly shared mailboxes, team calendars and archived discussion, moving it into Microsoft 365 removes the routing problem rather than repairing it, and Microsoft 365 Business Standard covers that for a smaller organisation. Treat it as a project with its own plan, not as a fix for today’s error, and we can help size the licences once you have decided the migration is worth doing.
Every code this article covers
| Code | What it points at | Source |
|---|---|---|
MapiExceptionNoReplicaAvailable |
The single public folder mailbox holding that folder’s content could not be reached. Microsoft documents that public folder content is not replicated across mailboxes, so there is no second copy | not published by the vendor |
MapiExceptionNotAuthorized |
The mailbox was reachable but the caller lacks rights on the folder | not published by the vendor |
MapiExceptionObjectNotFound |
The store could not find the folder or mailbox object the request named | not published by the vendor |
ErrorFolderNotFound |
Web services response: the folder ID specified does not correspond to a valid folder, or the delegate does not have permission to access the folder | Microsoft Learn |
Confirm the fix worked
- The failing folder opens from a client that previously could not.
Get-PublicFolderon that path names a content mailbox whose database is mounted.- At least one public folder mailbox is serving hierarchy with
IsExcludedFromServingHierarchyclear. Get-Mailbox <user> | Format-List *PublicFolderMailbox*shows an effective mailbox that exists and is healthy.- The application event log shows no new public folder errors after the change.
Questions people ask about this
Do modern public folders really have no replicas?
Correct. Microsoft states that content is stored in the public folder mailbox and is not replicated across multiple public folder mailboxes. Redundancy comes from database availability group copies of that mailbox’s database, not from folder replicas.
Why does restarting Outlook fix it sometimes?
Because the client caches its hierarchy binding and the folder tree. Once the server side is repaired, a cached client can keep failing until it reconnects and asks again.
Should I pin users to a public folder mailbox or leave it empty?
Leave it empty unless you have a reason. Users are distributed evenly across the available mailboxes by default, and a pinned user does not move when you add capacity or take a mailbox out of service.
How do I find which mailbox holds a folder’s content?
Run Get-PublicFolder against the folder path and read ContentMailboxName. That names the one mailbox that has to be healthy for that folder to open.
Does fixing this require any purchase?
No. Every step here is configuration. Licensing only enters the picture if you decide to retire the on-premises public folder estate altogether, which is a project rather than a fix.
