gstripe stripe size units?

[Pawel Jakub Dawidek]

On Fri, Sep 24, 2004 at 05:51:54PM -0400, Paul Mather wrote:
+> > If gstripe is running in "fast" mode (kern.geom.stripe.fast=1), size of
+> > stripe could be small, because then, it still sends as large I/O requests
+> > as possible and reorganize the data in memory, but this method consumes
+> > a lot of memory if you want it to be efficient.
+> 
+> One thing that puzzles me is that no matter how large I made the stripe
+> size, I never got a kern.geom.stripe.fast_failed > 0. [...]

It will be hard for raidtest tests, but it is really easy for
UFS+softupdates and small stripe size.

+> I'm puzzled because a stripe of 16 MB will not fit in 6553600 bytes, so
+> surely fast_failed should be > 0 at stripes of 8 MB or greater in my
+> tests for the above value of kern.geom.stripe.maxmem?

This is a problem for small stripe sizes, not for big ones.
Remember, that no metter how large stripe size is, maximum I/O request
size is still 128kB, so for stripe size >= 64kB there is not even chance
to fail - gstripe don't allocate memory if it is not needed, i.e. if we
can split request on two pieces or even send it as one piece.

+> Also, I don't know what the distribution of request sizes is in
+> raidtest.data.  The raidtest program operates over the raw device, and
+> so may not necessarily behave in terms of issuing requests as, say, a
+> UFS filesystem might.

Of course. You should run some file systems tests as well.

-- 
Pawel Jakub Dawidek                       http://www.FreeBSD.org
pjd at FreeBSD.org                           http://garage.freebsd.pl
FreeBSD committer                         Am I Evil? Yes, I Am!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-geom/attachments/20040925/c18bf2f8/attachment.bin