Michael Brutman wrote: > Somehow I expected a very quick reply from you. :-)
From me, or from the list? ;)
> My comments on kernel mode were based on profiling data that I collected.
> The profiling data showed the time to be in process/thread creation, not
> I/O.
How you profiled it? (just for me to know)
> I know that exim is nasty on the disks. I've run tests with EXT3 and
> Reiser, with mixed results. Both of those are journalling file systems so
> I expect there to be more writes than usual. What is recommended ? Does
> it make sense to run EXT2? (I plan to measure and find out.)
Ext2 will be faster compared to Ext3, i'm not a fan of ReiserFS, as it
was broken too often.
> On my system exim is generating about 1000 disk writes per second. The
Processing lot's of mails, eh? ;)
AFAIK you can gain HIGH improvements, by removing the fsyncs from the
exim-source, but if your system crashes, noone can garantuee for your
data (lost mails.).
> 1000 disk writes are spread across 6 disk drives. I'm on an iSeries
> running Linux, so the disk performance is hard to quantify - there is a lot
> of double buffering going on. Regardless, my measurements didn't point me
> at disk - it was CPU bound on my box, and most of the time was in the
> kernel. (Of which I think it was process creation, based on the profile.)
> Disk may be a problem, but I don't think it is one on this box yet.
Just send me one of this beasts, and i will make the tests ;)
Just a guess, but is it possible that the PowerPC-Arch is not very
good/fast in creating/forking processes.
If you tell us what you want to do, we could tell you how to optimize
the system for this case, there are many ways to tune exim in one or
another direction.
