Highlighted

Open Replicator scenario for how many hours of changed data is lost ??


Some one please explain this as i am not able to get it .

Question Image

Refer to the exhibit.

A customer is running a migration of several volumes from a DMX4 to a VMAX.
The following procedure shows the planned timeline and steps:

•00:00 hrs Timefinder/Snap established of DMX source volumes
•00:15 hrs Host IO stopped
•00:30 hrs Open Replicator Hot Pull session activated
•01:30 hrs Host IO started
•08:45 hrs Open Replicator session completed and deactivated
•09:00 hrs Timefinder/Snap terminated

As shown in the exhibit, the client follows the migration process. Data corruption is discovered by the team validating the hosts 1/2 hour after host IO is started. The Open Replicator session is stopped at time 02:30 and recovery is initiated.

How many hours of changed data is lost?

Tags (2)
0 Kudos
3 Replies
Frederik.K
1 Copper

Re: Open Replicator scenario for how many hours of changed data is lost ??

Hi,

There are not enough information in the text to dig more i guess...

But i would try simply : snap is taken at midnight. snap is restored at 2.30. There are potentially 2h30 hours lost.

By the way i would have done things differently in order to have a consistent snap (except if they managed it by the application):

•00:00 hrs Host IO stopped

•00:15 hrs Timefinder/Snap established of DMX source volumes

•00:30 hrs Open Replicator Hot Pull session activated

0 Kudos
sauravrohilla
3 Cadmium

Re: Open Replicator scenario for how many hours of changed data is lost ??

well, theoretically, there should be only 1:15 hours of data loss.

SNAP was activated at 00:00

Host IO stopped 00:15

Host IO restarted 01:30

SNAP restored 02:30

So, host was not writing for 1 hr 15 mins so the actual DL should be of 1.15 hrs

regards,

Saurabh

0 Kudos
Frederik.K
1 Copper

Re: Open Replicator scenario for how many hours of changed data is lost ??

To be fair, i said 2.30 hours because the Host IO stopped and Host IO restarted is never guaranted. If another host is writing, to the volume, if an error is made withint the array with the Open Replicator copy process, etc... so i agree with you in a perfect world, but this is not a perfect world...

0 Kudos