Re: kswapd and system response

Re: kswapd and system response

[Marcelo Tosatti]

On Sat, 24 Nov 2001, Peter Jay Salzman wrote:

> hi there,
> 
> every so often, my system (2.4.13 smp) gets really sloooooow.   a typical top
> looks something like:
> 
>   PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
>     5 root      14   0     0    0     0 RW   99.9  0.0   7:52 kswapd
>     7 root       9   0     0    0     0 SW    1.9  0.0   0:04 kupdated
>  2053 root      13   0   984  984   776 R     0.7  0.4   0:01 top
>  2063 p          9   0  2080 2080  1292 S     0.3  0.9   0:00 vim
> 
> kswapd is definitely not behaving correctly.
> 
> the thing that brought it on this time was gtv (an mpeg viewer which uses
> sdl) bailed on me.   X "kind of" froze, so i killed it.  but that's only the
> cause this time around.  last time, it happened between the time of going to
> bed and waking up (ie- the system was pretty much idling from a user's pov).
> 
> this has been happening with the 2.4.13 kernel every couple of days.
> 
> i thought i'd post it to lkml to let someone know that there's a problem.
> i'm not subscribed to the list, so if there's anything else i can do to help
> diagnose what's going on next time it happens, please cc me.
> 
> i feel like such a loser when i reboot.   ;)   i heard that 2.4.15 was just
> released, so i'll try upgrading to that and see if it helps any.

Don't use 2.4.15: it may corrupt your filesystem.

I just released 2.4.16-pre1 which has a fix for this problem...

Please try it and tell us if you still see the kswapd problems.

kswapd and system response

[Peter Jay Salzman]

hi there,

every so often, my system (2.4.13 smp) gets really sloooooow.   a typical top
looks something like:

  PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
    5 root      14   0     0    0     0 RW   99.9  0.0   7:52 kswapd
    7 root       9   0     0    0     0 SW    1.9  0.0   0:04 kupdated
 2053 root      13   0   984  984   776 R     0.7  0.4   0:01 top
 2063 p          9   0  2080 2080  1292 S     0.3  0.9   0:00 vim

kswapd is definitely not behaving correctly.

the thing that brought it on this time was gtv (an mpeg viewer which uses
sdl) bailed on me.   X "kind of" froze, so i killed it.  but that's only the
cause this time around.  last time, it happened between the time of going to
bed and waking up (ie- the system was pretty much idling from a user's pov).

this has been happening with the 2.4.13 kernel every couple of days.

i thought i'd post it to lkml to let someone know that there's a problem.
i'm not subscribed to the list, so if there's anything else i can do to help
diagnose what's going on next time it happens, please cc me.

i feel like such a loser when i reboot.   ;)   i heard that 2.4.15 was just
released, so i'll try upgrading to that and see if it helps any.


pete

-- 
PGP Fingerprint: B9F1 6CF3 47C4 7CD8 D33E  70A9 A3B9 1945 67EA 951D
PGP Public Key:  finger p@dirac.org

Re: kswapd and system response

[Ahmed Masud]

On Sat, 24 Nov 2001, Peter Jay Salzman wrote:

> hi there,
>
> every so often, my system (2.4.13 smp) gets really sloooooow.   a typical top
> looks something like:
>
>   PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
>     5 root      14   0     0    0     0 RW   99.9  0.0   7:52 kswapd
>     7 root       9   0     0    0     0 SW    1.9  0.0   0:04 kupdated
>  2053 root      13   0   984  984   776 R     0.7  0.4   0:01 top
>  2063 p          9   0  2080 2080  1292 S     0.3  0.9   0:00 vim
>
> kswapd is definitely not behaving correctly.
>
> the thing that brought it on this time was gtv (an mpeg viewer which uses
> sdl) bailed on me.   X "kind of" froze, so i killed it.  but that's only the
> cause this time around.  last time, it happened between the time of going to
> bed and waking up (ie- the system was pretty much idling from a user's pov).
>
> this has been happening with the 2.4.13 kernel every couple of days.
>

Hello

I have experienced similar behaviour with kswapd/kupdated with 2.4.12 and
2.4.13 .  After much struggle, I discovered that the system had become
slugish and non-responsive because i was using the wrong IDE busmastering
and DMA drivers (VIA chipset) instead of the correct ones for my
motherboard (PROMISE chipset). If you are using IDE, then perhaps you have
something similar and are using the incorrect drivers?

If you are indeed using IDE then suggest that you turn of busmastering all
together and give it another try.

Ahmed.

Re: kswapd and system response

[James Davies]

On Sun, 25 Nov 2001 06:48, Ahmed Masud wrote:
> On Sat, 24 Nov 2001, Peter Jay Salzman wrote:
> > hi there,
> >
> > every so often, my system (2.4.13 smp) gets really sloooooow.   a typical
 <snip>
> > this has been happening with the 2.4.13 kernel every couple of days.
>
> Hello
>
> I have experienced similar behaviour with kswapd/kupdated with 2.4.12 and
> 2.4.13 .  After much struggle, I discovered that the system had become
> slugish and non-responsive because i was using the wrong IDE busmastering
> and DMA drivers (VIA chipset) instead of the correct ones for my
> motherboard (PROMISE chipset). If you are using IDE, then perhaps you have
> something similar and are using the incorrect drivers?
>
> If you are indeed using IDE then suggest that you turn of busmastering all
> together and give it another try.
>
> Ahmed.

	I am having the same problem with 2.4.6. I have the VIA Drivers compiled in,
	which are the correct ones for my chipset. Perhaps it is a problem with the
	VIA Drivers? I havn't tested it with anything else. 
	


_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com

Re: kswapd and system response

[Ahmed Masud]

>
> 	I am having the same problem with 2.4.6. I have the VIA Drivers compiled in,
> 	which are the correct ones for my chipset. Perhaps it is a problem with the
> 	VIA Drivers? I havn't tested it with anything else.
>

Could very well be... VIA chipset has been a bit of a snag for me in the
past.

Ahmed.