Hans-Werner Hilse schreef:
> Hi,
> 
> On Mon, 31 Oct 2005 17:24:26 +0100 Holly Bostick <[EMAIL PROTECTED]> 
> wrote:
> 
> 
>> Basically, my system is running "fine" (no overt problems), but 
>> about every 30 seconds or so, it 'pauses' to do something, and I 
>> have to wait for 5-10 seconds while it does it before I can go 
>> further. Or the display pauses, and I have to wait for the redraw ,
>>  I can't tell which.
> 
> 
> It all gets down to the question whether it "stays" in Userland or in
>  Kernel context in such a situation. If this happens in Userland, CPU
>  utilization will probably be high in such a situation and it must be
>  due to a high priority process when X is that much slowed down - 
> which itself is already running at high prio.
> 
> So I very much guess it's a kernel issue. Compiling a fresh new one 
> would be my first advice.

Thanks for the tip; this is my first experience with 2.6.13, and I know
that it's not 'perfect' as yet. What was really suggesting to me that it
might be a kernel scheduling issue (aside from the fact that I know that
kernel scheduling issues are coming rapidly to a boil in several areas),
is that nothing has particularly changed on my system, inosfar as I have
run several high-priority processes concurrently under X with previous
kernels, but have not noticed this issue under, say, 2.6.12.

If this is a kenel issue, is 2.6.13-r5 likely to be any better than -r4?
I can't go to 2.6.14 since the ATI drivers won't compile under it, afaik
(that was the situation a couple of days ago, and I haven't checked to
see if the current 'unofficial' (?) patch has made it into any of the
trees yet). What could I do to minimize the problem (are any of the
schedulers set by default so "icky" that I should disable them, for
example)?
> 
> BTW, are you using any kernel based network filesystems (NFS, Samba, 
> ncpfs), FUSE, or PPP? The trick here would be to simply selectively 
> disable each one of these and watch if it keeps happening.
> 

Yes, I am using Samba, but it doesn't work all of a sudden (at least
from my end; it works from my bf's end, so I have to ask him to copy
files from my PC, because I can't reach his PC to do it myself;
connection times out. But he may have screwed up something in the
network, for all I know; I simply haven't had the time to check, though
the issue is moving slowly yet steadily up the chain of priority). But,
similar to my suspicion of gamin, it did occur to me that Samba could be
trying to connect behind the scenes, and repeatedly failing.

Anyway, I'm not going to be able to compile 2.6.13-r5 for a day or two
(well, I can compile it, but I won't be able to boot to it for a day or
two, so I'm not compiling it yet), so any other tips for special
treatment of the kernel or anything else are still more than welcome.

Holly
-- 
gentoo-user@gentoo.org mailing list

Reply via email to