4

I recently set up a three servers running MSSQL Server 2016 in a high safety synchronous mirroring configuration. Right now, everything is working normally. Any changes made in the principal are replicated on the mirror consistently.

However, both the Principal and the Mirror server constantly have the error:

"Database mirroring connection error 4 'An error occurred while receiving data: '10054(An existing connection was forcibly closed by the remote host.)'.' for 'TCP://WITNESS_SERVER:5022'."

The Witness server also has the exact same error toward both the Principal and the Mirror. There are no errors between the Principal and Mirror, and there are no other accompanying errors (Handshake rejected, etc...). Interestingly, these disconnects occur repeatedly exactly 100 seconds apart without cease. They do not happen at the same time for the Mirror and Principal, and currently are offset by about 12 seconds. Has anyone run into a similar issue or know exactly what sort of issue can lead to 100 second forced disconnects?

| improve this question | |
  • Generally, when you see An existing connection was forcibly closed by the remote host it's either an application killing the socket or a firewall killing the connection. I'd check the firewalls between the servers. – Sean Gallardy - Retired User Oct 27 '16 at 21:05
  • All three servers have the firewall disabled on private networks. In between disconnects, the Principal and Mirror both connect to the Witness and send and receive data as normal. This makes me think it probably isn't a firewall issue. If it was, wouldn't they be unlikely to connect and stay connected over 100 second time periods? – Rick Oct 28 '16 at 13:46
  • I wasn't talking windows firewall, I was speaking about other firewalls between the systems. The error states it was forcibly closed, which means something else did it. The fact it happens every 100 seconds tells me that there is something between them that checks for connections around that often and then stops traffic. Every time I had this happen it was a network appliance shutting down my sockets. – Sean Gallardy - Retired User Oct 28 '16 at 15:10
  • 1
    We are experiencing the same issue on SQL Server 2016. Did you find out about it? – Ola Hallengren Nov 16 '16 at 9:46
  • 1
    Possibly support.microsoft.com/en-au/help/3210699/… -- seemingly published today: try SP1 CU1 or RTM CU4 (CC @Ola) -- am testing now – jimbobmcgee Jan 18 '17 at 13:09
3

We reported the issue to Microsoft support and got it fixed. I have also tested SP1 CU1 and the database mirroring error messages have disappeared. So I think that all is good. https://support.microsoft.com/en-us/help/3210699/fix-error-messages-are-logged-if-database-mirroring-is-configured-by-transact-sql-in-sql-server-2016-and-no-database-activity-occurs-for-more-than-90-seconds

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.