No one had to sell their soul. We just re-wrote how snapshots are done in 6.x. This re-write changes how the blockmap file is generated and give lots of benfits.
1. REALLY fast snapshots, even on 1+ TB volumes! 2. 31 Snapshots per volume rather than 20. The blockmap file was the limiting factor 3. SIGNIFICANTLY less overhead for SnapMirror. The blockmap file of the volume must be moved over during a SnapMirror transfer. The old pre-6 blockmap file was 1/1024 of the volume size. So if you had a 1TB volume, your blockmap size was 1GB. This has been cut by I believe (don't quota me exactly on this) by a factor of 16 or 32, thus speeding up SnapMirrors as well, espcecially small incremental SnapMirror transfers.
You may have heard about the "Split Blockmap"? This is the re-write I describe above.
Enjoy!
-- Adam Fox NetApp Professional Services, NC adamfox@netapp.com
-----Original Message----- From: Todd C. Merrill [mailto:tmerrill@mathworks.com] Sent: Thursday, May 10, 2001 8:57 AM To: toasters@mathworks.com Subject: ONTAP 6.x snapshots
To balance my karma on this list, after yesterday's "floppy shuffle" thread....
# time rsh F840-filer-running-6.1 'snap create vol0 test' creating snapshot... 0.01u 0.02s 0:00.28 10.7%
28-hundredths of a second wall-clock time, including loading and executing rsh and building and taking down the TCP/IP connection?!
Whose soul had to be sold to make the new snapshot work so fast?
Until next time...
The Mathworks, Inc. 508-647-7000 x7792 3 Apple Hill Drive, Natick, MA 01760-2098 508-647-7001 FAX tmerrill@mathworks.com http://www.mathworks.com