Legacy HDS Forums

HRpM - Advanced Pair Operations for HUR+SI cascade

Discussion created by Legacy HDS Forums on Jul 5, 2012

Originally posted by: arappoport



Dear Community,

I would like to know how you are using HRpM (Replication Manager) if your replication is set up like this:

2 sites, PROD and DR.
From PROD, volume is HUR'ed to DR then cascaded with a SI pair:
PROD - DR
P --HUR--> S
.................P --SI--> S

let's assume that one DR scenario will see the active volume being the SI(S) volume. How would you failback to PROD HUR(P) via HRpM, knowing that the data must flow back from SI(S) to HUR(P)?

in diagrams, * being the active volume:

normal flow:
PROD - DR
P* --HUR--> S
...................P --SI--> S

DR declared, failover: (suspend)
PROD - DR
P //--HUR--// S
...................P //--SI--// S*

PROD ready, failback: (?)

step 1: (reverse resync)
PROD - DR
P <--HUR-- S
.................P <--SI-- S*

step 2: (suspend for end apps to move to PROD)
PROD - DR
P* //--HUR--// S
.....................P //--SI--// S

step 3: (normal flow)
PROD - DR
P* --HUR--> S
...................P --SI--> S

I am trying to see if we can avoid bypassing HRpM and breaking pairs.


thanks in advance for your input on this.

Kind regards,

Alex Rappoport

Outcomes