DR Recovery Back to Primary DC

Question

DR Recovery Back to Primary DC

Super Cat

SSCertifiable

Points: 7332
More actions
October 3, 2014 at 2:01 am

#293913

Scenario
AG failover to DR Node.
Setup
Server A – Primary Datacentre
Sever B – Primary Datacentre
Server Z – DR Datacentre
Primary DC down. Manual Failover to DR Datacentre.
Question.
At the point of failing back to the Primary when all is OK again. Are the changes made on the DR server DBs replicated to the Primary DC dbs during the manual failover back to the Primary DC.

Viewing 6 posts - 1 through 5 (of 5 total)

You must be logged in to reply to this topic. Login to reply

Andy Hyslop SSCrazy Eights Points: 9198 More actions · Answer 1

Talib123 (10/3/2014)
Scenario
AG failover to DR Node.
Setup
Server A – Primary Datacentre
Sever B – Primary Datacentre
Server Z – DR Datacentre
Primary DC down. Manual Failover to DR Datacentre.
Question.
At the point of failing back to the Primary when all is OK again. Are the changes made on the DR server DBs replicated to the Primary DC dbs during the manual failover back to the Primary DC.

Hi

Steps we have used in the past (rough outline)

Change Server Z and Server A to Synchronous Commit

Allow to Synchronize

Failover Z to A (with no data loss)

Change Server A and Server B to Synchronous Commit

This should allow a failover with no dataloss

Andy

==========================================================================================================================
A computer lets you make more mistakes faster than any invention in human history - with the possible exceptions of handguns and tequila. Mitch Ratcliffe

Super Cat SSCertifiable Points: 7332 More actions · Answer 2

Thanks, the full process can be found on page 25 of this doc.

http://msdn.microsoft.com/en-us/library/jj191711.aspx

You answered another question I was going to ask, about reverting back with no data loss.

Again Thanks

Perry Whittle SSC Guru Points: 233854 More actions · Answer 3

Talib123 (10/3/2014)
Scenario
AG failover to DR Node.
Setup
Server A – Primary Datacentre
Sever B – Primary Datacentre
Server Z – DR Datacentre
Primary DC down. Manual Failover to DR Datacentre.
Question.
At the point of failing back to the Primary when all is OK again. Are the changes made on the DR server DBs replicated to the Primary DC dbs during the manual failover back to the Primary DC.

Configure all 3 instances for synchronous commit with automatic failover between the nodes in primary site. If a failover to DR occurs as its synchronous data loss will be minimal if any at all. Failing back will not incur a data loss.

-----------------------------------------------------------------------------------------------------------

"Ya can't make an omelette without breaking just a few eggs" 😉

Andy Hyslop SSCrazy Eights Points: 9198 More actions · Answer 4

Perry Whittle (10/7/2014)
Talib123 (10/3/2014)
Scenario
AG failover to DR Node.
Setup
Server A – Primary Datacentre
Sever B – Primary Datacentre
Server Z – DR Datacentre
Primary DC down. Manual Failover to DR Datacentre.
Question.
At the point of failing back to the Primary when all is OK again. Are the changes made on the DR server DBs replicated to the Primary DC dbs during the manual failover back to the Primary DC.
Configure all 3 instances for synchronous commit with automatic failover between the nodes in primary site. If a failover to DR occurs as its synchronous data loss will be minimal if any at all. Failing back will not incur a data loss.

Its a good point Perry - depends on the bandwidth of your pipes however - if you have (and the business is willing to pay for) an extremely good connection no problem - if however the pipes between your DR box and the prod system are limited (as they usually are) setting all three to Sync commit could cause slowdowns

Just my 2 cents 🙂

==========================================================================================================================
A computer lets you make more mistakes faster than any invention in human history - with the possible exceptions of handguns and tequila. Mitch Ratcliffe

Perry Whittle SSC Guru Points: 233854 More actions · Answer 5

Andy Hyslop (10/7/2014)
Its a good point Perry - depends on the bandwidth of your pipes however - if you have (and the business is willing to pay for) an extremely good connection no problem - if however the pipes between your DR box and the prod system are limited (as they usually are) setting all three to Sync commit could cause slowdowns
Just my 2 cents 🙂

Sure that's completely valid, but if you go to all this trouble and skimp on the network then it was pointless in the first place. depends how busy the databases are too

-----------------------------------------------------------------------------------------------------------

"Ya can't make an omelette without breaking just a few eggs" 😉