drax
ADSM.ORG Member
We currently offsite our copy storage pools via a tape, courier & vault rotation.
I'm implementing a WAN based alternative using virtual volumes and it looks semi-straightforward.
I create a DR TSM Server instance (instance B) to manage the archived data from the prod TSM servers virtual volumes (instance A).
To implement DR I would restore and run instance A to run on the same server as instance B.
This is required as only instance A's database knows what data is in the virtual volumes stored on instance B.
(Sing out if I got that wrong please!)
The BIG question is how do I conduct a DR test without interrupting my normal DR offsiting?
For example, can I rebuild instance A "read-only" in DR on the same server as instance B while the real "instance A" in production is still running and contacting DR?
By read only I mean, "nomig, noexpire, noshed" and whatever other dsmserv.opt entries I need to avoid my DR instance A from doing anything but read from instance B vvols.
A DR client side change (dsm.opt) to target the DR instance A should help avoid attempts to contact the real thing.
How does your site handle DR testing involving a server used for DR replication?
I'm implementing a WAN based alternative using virtual volumes and it looks semi-straightforward.
I create a DR TSM Server instance (instance B) to manage the archived data from the prod TSM servers virtual volumes (instance A).
To implement DR I would restore and run instance A to run on the same server as instance B.
This is required as only instance A's database knows what data is in the virtual volumes stored on instance B.
(Sing out if I got that wrong please!)
The BIG question is how do I conduct a DR test without interrupting my normal DR offsiting?
For example, can I rebuild instance A "read-only" in DR on the same server as instance B while the real "instance A" in production is still running and contacting DR?
By read only I mean, "nomig, noexpire, noshed" and whatever other dsmserv.opt entries I need to avoid my DR instance A from doing anything but read from instance B vvols.
A DR client side change (dsm.opt) to target the DR instance A should help avoid attempts to contact the real thing.
How does your site handle DR testing involving a server used for DR replication?
