What should we do to alter virtual fire drill data

[Zahid_Haseeb]

Environment

SFHA/DR = 6.0.1

OS = Windows 2008R2

Clustered Application = SQL 2008R2

Query

At DR Site Service Group for FireDrill is created and online. Any thing I change at Primary SIte Active node, refelect the changes at DR Site fire drill snapshot volume (via opening the "SQl Server management studio console" at DR Site ). Till here every thing is fine. Now my question is if I want to alter the snapshot volume data(of fire drill service group), I should disconnect the volume(snapshot volume) from DR site read only volume(replicated volume) whom it(snapshot volume) is being syncing. Would anyone share quick steps to do it (via VEA) as well as attaching the same volume(snapshot volume) so that the volume(snapshot volume) can update again with lates data.

[Gaurav_S]

Hi Zahid,

First thing, why would you want to change the data on primary snapshot volume ? Firedrill is a virtual point in time test to find out the failover capability of your DR site. I am trying to understand here on what you are trying to achieve by altering data on primary snapshot volume.

From Disaster recovery solutions guide:

Fire drill service groups do not interact with outside clients or with other instances of resources, so they can safely come online even when the application service group is online on the primary site

In a VVR environment, when running the fire drill, the VVRSnap agent does the following:

a) Detaches the mirrors from the original volumes to create point-in-time snapshots of the production data

b)Creates a fire drill disk group on the secondary site with a snapshot of the application data to use for testing purposes

as you can see, point a) explains that it is a point in time snapshot of production data & if you alter the configuration in snapshot somehow, you are not doing the correct firedrill ..

G

[Zahid_Haseeb]

I want to use it for testing the data

[Gaurav_S]

Well in that case shouldn't it be good to simply keep the point in top snapshot which was created by firedrill ?that will actually demonstrate a real test case for VFD.

G

[mikebounds · Accepted solution]

I don't quite understand your post:

Suppose your data at primary site is called D1 and this is replicated to DR site so it's copy is D2 and you have a snap plex S2 at DR site.  Then when firedrill service group is offline, D1, D2 and S2 are all the same.

When you online the firedrill service group this takes a snapshot so then S2 snap plex is disconnected from D2 so that changes to D1 are still made to D2, but NOT to S2, hence I don't understand what you mean by "Any thing I change at Primary SIte Active node, refelect the changes at DR Site fire drill snapshot volume", as S2 should NOT be changing when you change D1.

When you bring up SQL in firedrill service group using S2, most likely SQL will write to S2 as it is starting and it is the norm, then to test by reading AND writing to SQL.  When you have finish testing, you offline firedrill service group and this will snapback the plex and sync the changes, so when you online the firedrill service again, any changes made on D1 since the last firedrill will now be in the S2 snapshot.

So there is no need to do anything in VEA - testing your data on your DR site is what firedrill service group is for, except you shouldn't test it for an extended period as while you testing it, you can't failover to DR.

Mike

[Zahid_Haseeb]

Excellent words mike. This is what I was asking

[mikebounds]

Just to add, by default a firedrill service group will be using the same IPs, same drive letters and same SQL instance name as the real one, so this is why you can't failover to DR while firedrill service group is online.

Mike