Unsolved
This post is more than 5 years old
14 Posts
0
1224
When NOT to use the -differential flag with symclone
Hi everyone,
We have 16 sets of disks that are ( at least theoretically ) available to be replicated on an any-to-any basis using symclone.
Some clone operations will be repeats and so we expect to use "symclone -differential create" followed some time later by "symclone recreate".
Some clone operations will require a target to subsequently become a source. Here we expect to use "symclone create" followed sometime later by "symclone terminate".
Would we incur any performance penalty or otherwise face operational difficulties if we always used the -differential flag with symclone create, even if we decided subsequently to terminate a session without using symclone recreate one or more times during the lifetime of the SRC-TGT relationships?
We know that -differential creates SDDF sessions - is this a possible overhead ?
Our scripting would be much easier if we just used the -differential flag regardless ...
Cheers,
Nick
We have 16 sets of disks that are ( at least theoretically ) available to be replicated on an any-to-any basis using symclone.
Some clone operations will be repeats and so we expect to use "symclone -differential create" followed some time later by "symclone recreate".
Some clone operations will require a target to subsequently become a source. Here we expect to use "symclone create" followed sometime later by "symclone terminate".
Would we incur any performance penalty or otherwise face operational difficulties if we always used the -differential flag with symclone create, even if we decided subsequently to terminate a session without using symclone recreate one or more times during the lifetime of the SRC-TGT relationships?
We know that -differential creates SDDF sessions - is this a possible overhead ?
Our scripting would be much easier if we just used the -differential flag regardless ...
Cheers,
Nick
dynamox
2 Intern
2 Intern
•
20.4K Posts
1
July 11th, 2007 14:00
rawstorage
419 Posts
1
July 17th, 2007 05:00
dynamox is correct here.. there is no associate overhead on the symmetrix with the SDDF session that is created for the clone session.
As stated once the bit is flipped it doesn't require any more effort on the symmetrixes behalf.