Subject: Re: explaining TOP memory output and constant 1.0 load averages
To: None <netbsd-users@netbsd.org>
From: Mark Cullen <mark.r.cullen@gmail.com>
List: netbsd-users
Date: 07/19/2006 12:09:12
Johnny Billquist wrote:
> Good points, and questions...
> At this time, I've basically boiled down my attitude that disk caches 
> are good, but I never want them to interfere with other types of memory 
> needs, so I basically holds the disk cache allowance very low.
> If there is memory to use, the disk cache will grab it anyway, so it's a 
> good deal in my eyes.
> 
> Others, I'm sure, have other opinions on this, and I'll let them speak 
> for themself.
> 
> My /etc/sysctl.conf looks like this, anyway...
> 
> ----
> # Tune the vm.
> vm.filemin=5
> vm.filemax=10
> vm.anonmin=5
> vm.anonmax=80
> vm.execmin=5
> vm.execmax=50
> vm.bufcache=5
> ----
> 
> While I probably could work some more on these values, this is what I 
> put together in a minute or five, and it gives me much better 
> performance than before. So good, in fact, that I don't feel the need to 
> tweak the memory system any more.
> 
>     Johnny
> 

Thanks Johnny.

I think my load average problem may be something to do with Samba or 
RAIDFrame?! That's the only thing I haven't really tried to kill (it'd 
be a bit of hassle because all of my computers are logged on to the 
server). I've tried killing everything but Samba and it doesn't make any 
difference.

I'm guessing that it's not really anything to be overly concerned about, 
as I am not experiencing slowness and the CPU time in `top` says it's 
nearly always idle. It's a little annoying having it constantly ~1.00 
when it's not actually doing anything most of the time, but never mind 
:-) Makes it look busier than it actually is I suppose!

-- 
Mark Cullen <mark.r.cullen@gmail.com>