Re: Fw: Performance drop 2.6.0-test7 -> 2.6.1-rc2

[Andrew Morton]

Jan Kasprzak <kas@xxxxxxxxxxxxxxxxxxx> wrote:
>
> Andrew Morton wrote:
> : Jan Kasprzak <kas@xxxxxxxxxxxxxxxxxxx> wrote:
> : >  - this is reliable: repeated boot back to 2.6.1-rc2 makes the problem
> : >  	appear again (high load, system slow has hell), booting back
> : >  	to -test7 makes it disappear.
> : 
> : Is the CPU load higher than normal?  Excluding I/O wait?
> 
> 	No, ~30% system is pretty standard for this server. I have looked
> just now (2.6.0-test7), and I have 33% system, about 50% nice,
> and the rest is user, iowait and idle. Under 2.6.1-rc2 it was about 30%
> system and the rest iowait, with small amount of nice and user.
> However, the load may be different. It is hard to have any kind of
> "fixed" load when you serve data over FTP, HTTP, rsync and do some
> other minor tasks (updatedb, current/up2date server, ...).

OK.

> 	Do you still want the system profiling info?

Nope.

It would be interesting to run some simple benchmarks (dbench, iozone,
tiobench, etc) on a relatively idle system.

After that, it'd be a matter of searching through kernel versions, which
you presumably cannot do.  Or eliminating device mapper from the picture,
which is also presumably not an option.


Have you run `hdparm' to check that all those disks have DMA enabled?  I
guess they have.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/