The documentation has diagrams with availability groups taking part in multiple Distributed Availability Groups (AG 1, AG 2 and AG 3) take part in Distibuted AG1 ([AG 1, AG 2]) and Distributed AG2 ([AG 2, AG 3]).

Is there a limit to this?

The documentation says "As long as two availability groups can communicate, you can configure a distributed availability group with them".

The documentation also says "In other words, a primary replica can participate in two different distributed availability groups"

Is 'two' the max? Or just used for the purpose of illustrating that there can be more than one. Can a primary replica participate in three different DAGs?

The book "Pro SQL Server Always On Availability Groups" says that an availability group can be a member of more than one distributed availability group, it does not explicitly say two. https://link.springer.com/book/10.1007%2F978-1-4842-2071-9

  • It's an interesting question. May I ask why you are asking? I'm curious how many you'd want? A distributed AG has a specific use case or set of use cases. Also - in a previous question you just asked you were asking about Basic Availability Groups - you can't do distributed with Basic AGs. – Mike Walsh May 26 '18 at 1:12
  • I'm doing something that probably requires more SQL expertise than I have. :) At a high level, I want to have a read/write HA instance of SQL in 1 datacenter. Asynchronously there is a second HA copy that is in another datacenter where a read-only workload takes place. Additionally there are 1 to 2 DR copies. Each DR copy is additionally in a separate datacenter. Ideally each DR copy is also HA, but that's not required. Until now I have only used Mirroring for HA. Now i'm looking at AG, BAG, DAG or possibly FCI for HA. Until now I've been paying Standard Edition licenses. – Marksl May 26 '18 at 4:53

You are misreading that document. In the document and image you are referring to - there are actually two distributed Availability Groups in the image you refer to. each one containing two Availability Groups.

A distributed Availability Group is distributed between two (and as far as I now - only two) Availability Groups.

You can have multiple secondaries in an Availability Group and they can span data centers, so depending on what you are trying to do, you may not need a Distributed Availability Group to do it.

  • Thank you. I see how my terminology is wrong. I've tried to update my question to better communicate what I'm trying to do. Can you chain pairs of AG's indefinitely in DAGs? If I had availabilty groups 1, 2, 3,n and n+1. Can I have DAGs - [AG1, AG2], [AG2, AG3], [AG3, AG(n)], [AG(n), AG(n+1)]? – Marksl May 26 '18 at 4:59

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.