May 25, 2015 at 5:23 pm
Currently i am testing the AOAG fail over on boxes to be deployed to production in future and guess what, its not going to be that smooth as i initially thought.
I have two servers in different data centres with AOAG configured and databases are in synchronised mode which shows that configurations are correct and AOAG is working. Quroum is configured on a network share on different server.
Now starts the fail over testing
1. Primary Node shutdown and AG fails to secondary node. This test was carried out on both servers individually on both servers - PASS
2. Manual fail over of AG from SSMS by connecting to secondary instance - PASS
3. On primary server, SQL Server service is stopped from Windows Service manager which takes Availability Group cluster resource offline. Application and users are unable to connect to primary instance. Unfortunately AG didn't fail over to secondary server and test failed - FAILED
Third test was to see if windows service crashes due to any reason.
The only logical reason i can think off for AG not failing over to secondary server is that server is still alive and holds the quorum vote. This will not trigger fail over of AG to secondary server.
As per link (https://msdn.microsoft.com/en-us/hh710061), AG should automatically fail over to secondary server with SQL Server service being down.
Am i missing any configuration to resolve this test case? Or it is meant to behave this way due to cluster properties (i doubt and hope i am wrong)?
May 26, 2015 at 12:44 am
found the solution to the problem
Cluster Availability Group Policy
May 26, 2015 at 6:38 am
Interesting! Thanks for the update. It might help somebody else.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply