Hi list,

I'm not a performace person, but I wrote a simplistic script to test file
system performance. This started with the original query of why are ext3
file systems so inefficient when populated with many small files.  It
tests ext3 file systems with 1K and 4K block sizes, reiser and JFS. It
also tests those four file systems on FCP disk, for a total of 8 file
systems, all around 400MB in size.  It tests creating files of 200, 2000,
20000 and 200000 bytes until the file systems are full. It then tests read
and write times.  I started to analyze the output and report on it but got
called away to Other Things.

Mark Post was kind enough to put the partial report on the Web at
http://linuxvm.org/Info/fstest.pdf

Have a look if you are so inclined.  Feedback off-list will be
appreciated.

And don't forget Jens Osterkamp (who is a performance person) sent his
"Linux filesystems" SHARE pitch to Mark which is at
http://linuxvm.org/present/SHARE102/s9257jo.pdf
There is *a lot* more analyzed, distilled data in this presentation.

-Mike MacIsaac, IBM  mikemac at us.ibm.com   (845) 433-7061


----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to