Re: wierd behaviour...
Michael O'Reilly (michael@metal.iinet.net.au)
Fri, 28 Feb 1997 17:34:09 +0800
In message <199702280825.QAA19101@metal.iinet.net.au>, Kevin Littlejohn writes:
> Better option might be to put the "no_atime" and "inum" patches in place
> (http://www.ecsnet.com/) - these produced a definate performance
> improvement for our news server.
>
> KevinL (now why didn't I just phone you with that info? :)
> Oh, btw:
>
> root 1 0.5 0.1 832 252 ? S Feb 26 14:23 init
> news 376 3.1 9.9 13700 12748 ? S Feb 26 82:51 /usr/sbin/innd -p4
Yes, I'm running those on the news spool, but this machine also has
shell users, and does a fair bit of mail. On the news only machine,
innd is higher than init, but when there's a process mix, I think init
is getting kicked a lot more frequently, and that's generating the
sync overheads.
Note that sync() on the original machine can take up to 20 seconds to
run (the machine normally run with a 30 - 50 meg disk cache).
Michael.