Hello All,
let me start by thanking Lydia Zhang and Shashank for your previous replies to my first posting on the forum.
Very Helpful replies!
Now I have another question and this one will be slightly more comprehensive…
I tend to be rather ‘long winded’ so sorry for the length of the question(s) but better a few more seconds of reading than back and forth trying to clarify, right?...it might be a few more second though.
I will start with what I have and then with what I want.
Here goes:
I am working with a group that is using Maximo for a front end application and SQL Server 2012 as a backend.
They have Windows Server 2008 for the SQL Server 2012. There is no clustering, no always on failover, nothing but a stand alone sql server instance that they back up frequently.
On Thursday nights they shut down the Maximo App server and run maintenance on the SQL Server Instance.
That consists of DBCC Check DB, rebuild Index and update statistics.
From start to finish they are offline from Maximo from 5-7 hours.
So, that is what I have.
Now to what I want;
I am suggesting to the powers that be that they give me two (2) windows server 2012 boxes. Install either SQL Server 2012 or 2014 on both.
Move the Maximo Database to the first of the two SQL Server instances (let’s call that R1 for Replica 1) and then create an Always On Group for the Maximo Database and point that group to R2 (Replica 2) on the second Windows 2012 Box.
(Both Windows 2012 Servers obviously clustered).
The ‘what I want” described so far is ‘do-able’, I have done it before at least in a testing / learning project.
What I am fuzzy on but will need to be done is configuring a listener.
But for now let me put listener on the back burner and ask a few questions.
- If I have the setup like I just described, when that Thursday maintenance window starts is it possible to
suspend data movement and force failover from R1 to R2 and allow the users to continue working on Maximo pointed to R2 while I do the DBCC Check DB, rebuild Index and update statistics on R1? - If the answer to the above question is ‘yes’ then let’s say after the 5-7 hours and the Maintenance is over, I then resume data movement and fail
it back over to R1 from R2.
When I do that will R2 ‘synch’ up with R1 and move all the data changes for that 5-7 hours back over to R1?
Last but not least and I’m not going to consider this question # 3 as this I am still trying to get my head around this concept.
From my previous example, let’s call the Maximo database ‘MaxDB’.
I back it up off its current server and I am going to restore on R1.
Restore it to R1 and have the data file on F:\DATA and log file on G:\LOG.
When I set up the availability group wizard for MaxDB it will back it up and restore it to R2 but asks for a share folder that both R1 and R2 has read/write options to.
That folder is on R2 (because it has more drives and space) and the folder is called H:\AVAIL_SHARE
When MaxDB is restore to R2 it will have same data and log location as R1
(F:\DATA and log file on G:\LOG)
However when I go to H:\AVAIL_SHARE this only contains .bak files.
I have read that an availability group like I am describing ‘shares’ one transaction log?
Would that be the transaction long from R1?
I guess I will end it with that for now.
Any and all help will be much appreciated….thanks in advance
R. Brown