IDMS Backup Question Flash Copy

Discussion created by Gary_Cherlet on Apr 12, 2011
Latest reply on Apr 18, 2012 by Gary_Cherlet
I don't have write access to the list ( our company is read only). I singled a couple of you out, hope you don't mind.
How do you do back ups? We are a 24X7 shop. Currently each night we queisces are cv's for a split second, we then restart processing along with the hot backups. If we have a disaster we restore to that point in time. We don't rollforward because we have too many journals too much data. We have the potential to lose up to 23+ hours of processing depending on when the outage occurs..
We are now implementing a new disaster recovery process where in theory, we just IPL the DR site and bring up IDMS. Idms will warmstart as if it were coming up from a hard CV / Mainframe crash. There is a tech note TEC477658 that says flashcopy can not guarantee integrity. (TEC477658 IBM Flash Copy Consistency Group can not guarantee integrity?)
This is how Global copy, flash copy, global mirroring, should work at our company.
Disk mirroring takes place constantly. A to B & B to C, B contains data that is still in-flight, C is the check point (or consistancy) spot. If A goes down, the pair is broken. A DR is called and C (point-in-time) is flashed over to B, then B copied to D. The DR box will be IPL'd using D. We should be able to restart IDMS at the DR site as if you were coming up from any CV / mainframe crash at our regular production data center. Ca's stance is this may not work all the time. Our tests are baring that out in that onetime out of 10 we ended up with a broken chain. The other tests may have had one too, we just didn't hit any. Do you have any experience with this? thanks Mike
[font=Courier New]SCHUMAN..............ST PAUL
Center (prod)........(DR - site)
8700 DISK............8700
A ...................B..C(X)..D
\____>>>>>>>>>>>_____/>>>\>>>/ [font]

From: Olmstead, Mike []