[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: htree and nfs benchmarks



Theodore Ts'o wrote:

On Wed, Jun 18, 2003 at 08:21:46PM -0600, kwijibo zianet com wrote:


It looks like htree improves performance when the storage device is
local (as per my previous post) but it looks like htree degrades
performance when used in conjunction with NFS. I am going to try it
with gigabit ethernet when I get the time to see if maybe I am
possibly hitting some sort of weird network bottleneck.



That's certainly a surprising result.


Any chance you could get an ethereal capture of the traffic between
the NFS client and server, both with and without htree?  It might shed
some light on what's going on, and perhaps suggest some way we can
either fix the NFS server, or the interaction between the NFS server
and the ext3 filesystem.

- Ted


_______________________________________________ Ext3-users mailing list Ext3-users redhat com https://www.redhat.com/mailman/listinfo/ext3-users




Yeah, I was a bit surprised as well, I ran one of the
tests again cause I thought I screwed it up.  Can someone
else do this to confirm what I am seeing?

You want the full packets or just the headers?  It is going
to be a big chunk of data with the full packets.  It takes
about an hour or so for Postmark to run with the settings
I have it at.  I can reduce the number of files to possibly
speed things up and produce a smaller dump.  About where is
the threshold on where normal ext3 starts to crawl with large
numbers of files in a directory?

Steve




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]