February 28, 2017 at 3:56 pm
I have Cluster1 with 2 nodes and a whole different Cluster2 with 2 nodes. Can I set up AlwaysOn between Cluster 1 to Cluster 2? Or this option only available in SQL 2016 (Distributed Availbility Groups)? If it is possible can you lead me to some documentation on how to set this up? Appreciate it.
February 28, 2017 at 4:20 pm
JP10 - Tuesday, February 28, 2017 3:56 PMI have Cluster1 with 2 nodes and a whole different Cluster2 with 2 nodes. Can I set up AlwaysOn between Cluster 1 to Cluster 2? Or this option only available in SQL 2016 (Distributed Availbility Groups)? If it is possible can you lead me to some documentation on how to set this up? Appreciate it.
Yes (https://msdn.microsoft.com/en-us/library/ff929171.aspx). I note that just because you CAN do something does not mean you SHOULD. :hehe:
Best,
Kevin G. Boles
SQL Server Consultant
SQL MVP 2007-2012
TheSQLGuru on googles mail service
February 28, 2017 at 4:38 pm
I'm not sure were on the same page. This article talks about how you set up Availability Groups with a single Cluster. I'm wanting to know we set up Availability groups in one cluster and then use a second cluster as the secondary replica to fail over too.
March 1, 2017 at 3:58 am
Can you have an availability group across two different failover clusters? Yes.
The issue is, generally, AG is meant as a replacement for failover clusters. Unlike failover clusters, there is no shared storage in AG. You have two independent servers with independent storage and you can fail between them as needed. If you set up AG between two sets of failover clusters, you will have to do two things. First, AG will be async only. You can't have a synchronous connection in that environment. Second, you can't have automatic failovers of the AG. Manual only is allowed. This is because you can't bypass the failover cluster and it has it's own failover process that your AG must work within.
So you can do this, but generally, you shouldn't.
"The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
- Theodore Roosevelt
Author of:
SQL Server Execution Plans
SQL Server Query Performance Tuning
March 1, 2017 at 4:17 am
Grant Fritchey - Wednesday, March 1, 2017 3:58 AMCan you have an availability group across two different failover clusters? Yes.The issue is, generally, AG is meant as a replacement for failover clusters. Unlike failover clusters, there is no shared storage in AG. You have two independent servers with independent storage and you can fail between them as needed. If you set up AG between two sets of failover clusters, you will have to do two things. First, AG will be async only. You can't have a synchronous connection in that environment. Second, you can't have automatic failovers of the AG. Manual only is allowed. This is because you can't bypass the failover cluster and it has it's own failover process that your AG must work within.
So you can do this, but generally, you shouldn't.
An availability group spanning 2 separate clusters is a temporary solution, this is because a resource in one cluster cannot be failed over to another cluster, the clusters do not share resources.
Cross cluster AGs and workgroup(domainless) clusters are supported under Windows 2016\SQL Server 2016.
JP10 - Tuesday, February 28, 2017 3:56 PM
I have Cluster1 with 2 nodes and a whole different Cluster2 with 2 nodes. Can I set up AlwaysOn between Cluster 1 to Cluster 2? Or this option only available in SQL 2016 (Distributed Availbility Groups)? If it is possible can you lead me to some documentation on how to set this up? Appreciate it.
Please read my stairway to AlwaysOn at this link
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
March 1, 2017 at 10:23 am
Grant Fritchey - Wednesday, March 1, 2017 3:58 AMCan you have an availability group across two different failover clusters? Yes.The issue is, generally, AG is meant as a replacement for failover clusters. Unlike failover clusters, there is no shared storage in AG. You have two independent servers with independent storage and you can fail between them as needed. If you set up AG between two sets of failover clusters, you will have to do two things. First, AG will be async only. You can't have a synchronous connection in that environment. Second, you can't have automatic failovers of the AG. Manual only is allowed. This is because you can't bypass the failover cluster and it has it's own failover process that your AG must work within.
So you can do this, but generally, you shouldn't.
Hello Grant, Do you have any documentation that gives step by step on how to do this? Or do we set it up the same way as if we were going to set up one cluster as the first replica and then to a stand alone as replica 2? Appreciate the help.
March 2, 2017 at 2:09 am
JP10 - Wednesday, March 1, 2017 10:23 AMHello Grant, Do you have any documentation that gives step by step on how to do this? Or do we set it up the same way as if we were going to set up one cluster as the first replica and then to a stand alone as replica 2? Appreciate the help.
Nope.
It's really not a good choice. In terms of setup, you're setting up async, other than that, it's basically the same. However, for details, I'd strongly suggest reading Perry's articles.
"The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
- Theodore Roosevelt
Author of:
SQL Server Execution Plans
SQL Server Query Performance Tuning
March 2, 2017 at 6:49 am
JP10 - Wednesday, March 1, 2017 10:23 AMGrant Fritchey - Wednesday, March 1, 2017 3:58 AMCan you have an availability group across two different failover clusters? Yes.The issue is, generally, AG is meant as a replacement for failover clusters. Unlike failover clusters, there is no shared storage in AG. You have two independent servers with independent storage and you can fail between them as needed. If you set up AG between two sets of failover clusters, you will have to do two things. First, AG will be async only. You can't have a synchronous connection in that environment. Second, you can't have automatic failovers of the AG. Manual only is allowed. This is because you can't bypass the failover cluster and it has it's own failover process that your AG must work within.
So you can do this, but generally, you shouldn't.
Hello Grant, Do you have any documentation that gives step by step on how to do this? Or do we set it up the same way as if we were going to set up one cluster as the first replica and then to a stand alone as replica 2? Appreciate the help.
i've linked my stairway series on this site above
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
March 3, 2017 at 6:42 am
I have worked in an environment that was setup in the way JP10 is describing.
2 FCI clusters + STANDALONE DR server.
Each FCI cluster having a AOAG replica.
The FCI was used for HA
The secondary replica on the second Cluster was used as the read only copy.
DR server for DR and it worked well.
From a DBA point of view it was easy to configure, the hard part is the WINDOWS config. Which was done by another team.
March 3, 2017 at 1:03 pm
Which version of SQL? So your saying you have had replica1 (being a 2 node cluster) point to replica2 (being a different 2 node cluster)? So in this situation we have 2 different WSFC. My question if you say you can do this is how do you add replica2 when it's in a different WSFC?
Viewing 10 posts - 1 through 9 (of 9 total)
You must be logged in to reply to this topic. Login to reply