Update October 16, 2026: For more information related to installation of October 2026 updates for Windows Server and the possible impact on Exchange Server, please see Exchange servers may become uns...
Updated Oct 16, 2026
Version 7.0
Blog Post
We are experiencing the same issue on three independent Exchange Server SE environments after installing KB5128895 (Build 15.2.2562.56).
After approximately 24–36 hours, mailbox access fails with MapiExceptionMdbOffline (ec=1142), although all databases remain mounted and the Exchange services are running.
Outlook and OWA stop working, and mail delivery queues enter Retry state.
A full server reboot temporarily resolves the issue.
The affected systems run Windows Server 2019 and 2025.
Has the Exchange Team acknowledged this issue or identified a workaround?
- Nino_Bilic
Microsoft
I just added a banner on top of the blog post indicating there is an issue after installing Windows updates on Windows Server. We do not have a workaround as of now, the issue is still under investigation.
We are experiencing the same problems with a Microsoft Exchange 2019 server after installing KB5126651. Could we uninstall the update safely and will this solve the issue?
Nobody knows because Microsoft hasnt said anything, lol
Additional diagnostic findings from our eight independent Exchange Server SE environments.
We manage eight Exchange Server SE servers running Windows Server 2019 and Windows Server 2025. All eight servers experienced the same MapiExceptionMdbOffline error after installing the October 2026 Windows updates.
Affected Windows updates:
Windows Server 2019: KB5126651
Windows Server 2025: KB5126646- Consistent failure timing
Five independent servers were found to have failed at the following system uptimes:
22 hours 8 minutes
22 hours 17 minutes
22 hours 19 minutes
22 hours 40 minutes
22 hours 45 minutesOn another server, the issue occurred a second time after a complete reboot without changing the installed updates.
Last boot: October 15, 2026 at 08:46:29
First recorded Event ID 1006: October 16, 2026 at 06:45:55Uptime to first recorded error: 21 hours 59 minutes 25 seconds.
On one server, all three MAPI connectivity tests were successful at 22 hours 1 minute 19 seconds of uptime and failed at 22 hours 2 minutes 28 seconds.
The transition was observed within 69 seconds.
The timestamps represent when failures were detected, not necessarily the exact beginning of the underlying issue.
- Exchange processes remained running
We verified that Microsoft.Exchange.Store.Service, Microsoft.Exchange.Store.Worker and Microsoft.Exchange.RpcClientAccess.Service retained their original process IDs and start times before and after the failure.
All three mailbox databases lost MAPI connectivity simultaneously.
- Recovery tests
Restarting MSExchangeRPC did not resolve the issue.
Restarting MSExchangeIS did not resolve the issue. All three databases were dismounted afterwards.
A complete Windows Server reboot restored normal operation. All databases mounted successfully, all MAPI tests passed, and transport queues cleared.
Based on our tests, restarting individual Exchange services is not an effective workaround.
- Conclusion
We have observed the same failure pattern across eight independent Exchange Server SE installations.
The consistent timing around 22 hours of system uptime suggests a possible time-dependent mechanism within Windows Server or its interaction with Exchange.
We hope these findings help the Exchange and Windows Server teams identify the root cause.
Thank you for investigating this issue.
My previous comment is probably still pending approval, so I’m posting this again here: we are seeing the same issues in our environment, but we only installed the October Windows updates (KB5126646, KB5126645 and KB5127690) on Windows Server 2025, not the October Exchange SU (KB5128895). Our Exchange Server SE RTM DAG is still on September 2026 SU v2 (15.2.2562.53).
Thank you, this is very important information.
We have two affected Windows Server 2025 systems with exactly the same Windows updates installed:
- KB5126645
- KB5126646
- KB5127690
Both systems experienced MapiExceptionMdbOffline (ec=1142) approximately 24–36 hours after the updates.
Unlike your environment, we also installed Exchange SE October SU KB5128895.
Since your environment is still running Exchange SE build 15.2.2562.53, this strongly suggests that the Exchange October SU itself may not be required to trigger the issue.
Could you confirm whether you are seeing the same MapiExceptionMdbOffline error (ec=1142), with mailbox databases still showing as mounted?
This could help narrow down the root cause to the October Windows updates or an interaction with Exchange SE.
Yes, I can confirm. We see exactly the same error, and the databases were still showing as mounted and healthy at the time:
Event ID 1006 (MSExchange Mailbox Replication)
The Microsoft Exchange Mailbox Replication service was unable to process jobs in a mailbox database.
Database: DB01
Error: MapiExceptionMdbOffline: Unable to make connection to the server. (hr=0x80004005, ec=1142)
ECP was still working, and both ECP and the Exchange Management Shell showed the databases as mounted and healthy. However, we could not move them (switchover failed), and the Exchange services could not access them. OWA was down and the transport queues were filling up. Only a reboot resolved it.
The timing matches yours as well. The updates were installed during the night of October 13, and the problems started on the evening of October 14, roughly a day later.
So we have the same three Windows updates (KB5126645, KB5126646, KB5127690) and the same error, with and without KB5128895. That points pretty clearly to the October Windows updates, or an interaction between them and Exchange SE, rather than the Exchange October SU itself.
Did you install the October Windows and .NET updates at the same time?
Windows Server 2019, Updates were installed with WSUS
- Security Update For Exchange Server Subscription Edition (KB5128895)
- 2026-10 Cumulative Update for .NET Framework 3.5, 4.7.2 and 4.8 for Windows Server 2019 for x64 (KB5128923)
- 2026-10 Cumulative Update for Windows Server 2019 for x64-based Systems (KB5126651)
Thanks for your reply.
We compared all three affected Exchange Server SE environments.
All three servers are running Exchange SE with October SU KB5128895 (Build 15.2.2562.56) and experienced the same MapiExceptionMdbOffline issue approximately 24–36 hours after the updates.
The installed Windows updates and .NET Framework versions differ:
Server A:
- KB5126651, KB5126650, KB5127685
- .NET Framework 4.8, Release 528049
Server B and Server C:
- KB5126645, KB5126646, KB5127690
- .NET Framework 4.8, Release 533509
On Server C, the Exchange security update was installed first, followed by a reboot. The Windows updates were installed afterwards.
All three servers recovered after a full reboot.
The only identified update common to all three environments is Exchange KB5128895.
This suggests a possible regression in the Exchange security update, although an interaction with the October Windows updates cannot yet be ruled out.
For now, we are still on the September updates, with the October update for Exchange, and the issue does not seem to have occurred yet.
Exchange SE (october security update) on Windows Server 2025 (.Net 4.8.1 september security update).
In my case: Yes
For now, we are still on the September updates, with the October update for Exchange, and the issue does not seem to have occurred yet.
Exchange SE (october security update) on Windows Server 2025 (.Net 4.8.1 september security update).
