August 4, 2022 at 11:18 am
Hi guys,
Testing our AG failover we noticed that we get an failure when moving to any node other that Node1 of our cluster. It is failing to bring the IP of the AG online due to a conflict.
When looking at the ipconfig output I can see that the IP for the AG Role returns for ALL nodes. Its like its bound to each adapter.
I tested with the core cluster group also and this works fine. Its IP only appears on the ipconfig output for the node which it is currently living.
For reference we are cloud based IAAS using load balancers in case that matters
Any help would be greatly appreciated
Thanks,
V
August 5, 2022 at 12:10 pm
Thanks for posting your issue and hopefully someone will answer soon.
This is an automated bump to increase visibility of your question.
August 9, 2022 at 9:30 am
This was removed by the editor as SPAM
August 12, 2022 at 6:02 am
Your AG listener needs to have a unique IP address. It should not try to share the address with a node or any other object.
Your client applications connect to the listener name and this routes the SQL statements to the appropriate server. The IP address of the listener stays the same regardless of the node it is pointing to.
I may be wrong but it sounds like you are trying to use the LB to route connections to one of the nodes and also using the listener to do the same thing. This seems designed to produce conflicts.
Your LB should therefore be based around the Listener and does not need to know anything about the underlying infrastructure.
Original author: https://github.com/SQL-FineBuild/Common/wiki/ 1-click install and best practice configuration of SQL Server 2019, 2017 2016, 2014, 2012, 2008 R2, 2008 and 2005.
When I give food to the poor they call me a saint. When I ask why they are poor they call me a communist - Archbishop Hélder Câmara
August 12, 2022 at 9:38 am
Hi EdVassie,
Appreciate the response.
Configuration wise we were fine. The LB and the AG were correctly set up with individual unique IP. The problem we had was windows bound this IP to the local Network Adapter of each node. Hence every time we failed out the IP was already allocated to the local network adapter. This might have happened on Build or during some PS scripts to allocate Health Probes we are not sure.
The fix was to delete the IP from the network adapters on the secondary replica nodes with Netsh.
Once we did this we could failover successfully and the IP for the AG/LB was only present on the active node.
Thanks,
V
August 16, 2022 at 4:39 am
This was removed by the editor as SPAM
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply