[PVFS2-developers] metadata record pre-allocation

[Nathan Poznick]

Thanks to google, I found this thread:
http://www.uwsg.iu.edu/hypermail/linux/kernel/0207.2/0082.html
which seems to indicate that the "fsync() does the same thing as sync()"
behavior is known, and is rooted in the semantics of how data=ordered
journaling works.

It would be interesting to do some of the "real-life" PVFS2 metadata
tests with the meta server's storage space on say, ext2.

-- 
Nathan Poznick <poznick at conwaycorp.net>

Don't take up a man's time talking about the smartness of your
children; he wants to talk to you about the smartness of his children.
- Ed Howe

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://www.beowulf-underground.org/pipermail/pvfs2-developers/attachments/20040727/8bb2ded8/attachment.bin