Unsolved
This post is more than 5 years old
36 Posts
0
639
May 14th, 2015 05:00
Timefinder snaphosts consistency for rollback's
Hello Guys,
I have a doubt that using Timefinder Snapshots feature in VMAX.
How much granular consistency, we can ensure for our production databases.
And how can we ensure the how snapshot restores can help to fix the database transaction level corruption and how much granular can it be ?
Also, for a very mission critical env. how many snapshots are recommended in day for production luns ?
Please share your expert advice on the same.
No Events found!
rawstorage
3 Apprentice
•
420 Posts
1
May 14th, 2015 06:00
Anil,
Using the -consistent option on a snapshot activate creates a crash consistent copy of the data. A short hold is put on the volumes involved in taking the snap shot to ensure a consistent point in time. This creates a snapshot in the array that is write order consistent and would be considered crash consistent by a database or application.
Depending on the snapshot technology in the array you can be very granular, in VMAX3 you can take up to 256 snapvx snapshots of a production lun, that's enough for one every 6 minutes, these are consistent by default and do not require a target device so are space efficient and cache efficient.
For a mission critical application the decision is up to your business to decide how often to take them, as this will give you a recovery point objective