[Ocfs2-tools-commits] tag, ocfs2-tools-1.4.3, created. ocfs2-tools-1.4.0-361-gccdb324
svn-commits at oss.oracle.com
svn-commits at oss.oracle.com
Fri Aug 7 16:04:12 PDT 2009
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "Tools to manage the ocfs2 filesystem.".
The tag, ocfs2-tools-1.4.3 has been created
at ccdb3246ee30476b493bdfcdc201bbae88f06a62 (commit)
- Log -----------------------------------------------------------------
commit ccdb3246ee30476b493bdfcdc201bbae88f06a62
Author: Joel Becker <joel.becker at oracle.com>
Date: Thu Aug 6 17:54:00 2009 -0700
debugfs.ocfs2: Add a fragmentation score to the 'frag' command.
Fragmentation is often looked at to see the effect on I/O. But a 100%
fragmented file of 1MB clusters will perform a lot better than a 10%
fragmented file of 4K clusters.
This adds a 'fragmentation score' to the 'frag' command of
debugfs.ocfs2. The score is simply calculated as the fragmentation
percentage times the number of clusters per megabyte. Why a megabyte?
Partly because it's our largest cluster size, but also because we know
1MB I/Os are a nice sweet spot.
This gives us an easy way to compare the I/O of files across cluster
sizes. The higher the score, the worse the I/O.
A file with one extent has a score of 0. Our theoretical 100%
fragmented file of 1MB clusters has a score of 100. The 10% fragmented
file of 4K clusters has a score of 2560. Ouch! That 10% fragmented
file has an average extent size of 40K.
Signed-off-by: Joel Becker <joel.becker at oracle.com>
Signed-off-by: Mark Fasheh <mfasheh at suse.com>
-----------------------------------------------------------------------
hooks/post-receive
--
Tools to manage the ocfs2 filesystem.
More information about the Ocfs2-tools-commits
mailing list