September 29, 2015 at 3:00 pm
We have a sql 2008r2 active/passive cluster with a log ship RO node also part of solution that we want to move to sql 2014 and AG. THis supports a large 3rd party medical records database. One of two choices
Build new active/passive 2014 cluster with sql 2014 AG and also have a Read Only node for reports in somewhat isolated environment, migrate Database to new group, remove old active/passive cluster and move new AG environment into that domain keeping the name the same for primary. Not a big fan of this but it does save us from all the Citrix and other connections being changed.
2nd solution is build everything the same in current domain with new cluster name, shutdown old cluster after migrating database and make all necessary connections changes to Citrix, this would be our only current cluster and not sure solution one is worth the risk.
Got 2 months to do this so any opinions welcomed
September 30, 2015 at 4:38 am
Make no mistake this is an extremely involved task and will need careful planning. Couple of questions first though before I make a recommendation
so the clustered instance, is it a default or named instance?
how easy is it for you to change all of the reporting connections?
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
September 30, 2015 at 6:23 am
it is named, I am a little wary of building an AG and then bringing it in to a domain. DId this at old job with a 2008 cluster and we had issues. I would rather live with the pain point of changing connections, also we could stand up the ag and do some initial testing for connections. This is our first AG and 2014 server, trying to avoid any unforeseen complications, I can live with fixing connections
September 30, 2015 at 10:10 am
As it's a named instance, let's assume you have
THENETWORKNAME\Instance
You want a new cluster with an availability group and all clients to connect for write and readintent connections.
Dump the LS secondary server and use a readonly replica in the Availability Group to provide read intent query access. This will require the correct configuration of your read only routing scripts.
So, with that in mind, at a high level this is what I would do
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
September 30, 2015 at 11:01 am
that's kind of what my game plan was. When we made a cluster with the same name and then moved it into the domain the one it was replacing we had all kinds of trouble. Been a dba for 20 years but not a network expert and there were all kinds of klugy things going on. I believe the KISS principle is better served
October 1, 2015 at 6:43 am
tcronin 95651 (9/30/2015)
When we made a cluster with the same name and then moved it into the domain the one it was replacing we had all kinds of trouble.
Yep, you don't give it the same name. In fact the client access point you use for the cluster itself has no bearing on any instances installed. It's just the management entry point for the cluster.
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply