Box A - 750 MHz Athlon, 256M RAM, SuSE 6.4, 2.2.16
Box B - 850 MHz Athlon, 512M RAM, RH 6.2, 2.2.17
Here is the first few lines of top output from Box A
11:39pm up 3:10, 1 user, load average: 0.06, 0.03, 0.00
78 processes: 76 sleeping, 2 running, 0 zombie, 0 stopped
CPU states: 2.9% user, 0.5% system, 0.0% nice, 96.4% idle
Mem: 257472K av, 168856K used, 88616K free, 81212K shrd, 54604K buff
Swap: 136512K av, 0K used, 136512K free 44408K cached
PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND
317 root 16 0 62864 61M 2088 R 0 2.3 24.4 0:47 X
1509 niall 5 0 1052 1052 856 R 0 0.7 0.4 0:06 top
1484 niall 2 0 4152 4152 2948 S 0 0.3 1.6 0:01 konsole
1 root 0 0 200 200 172 S 0 0.0 0.0 0:05 init
2 root 0 0 0 0 0 SW 0 0.0 0.0 0:00 kflushd
3 root 0 0 0 0 0 SW 0 0.0 0.0 0:00 kupdate
4 root 0 0 0 0 0 SW 0 0.0 0.0 0:00 kpiod
5 root 0 0 0 0 0 SW 0 0.0 0.0 0:00 kswapd
21 root 0 0 0 0 0 SW 0 0.0 0.0 0:00 kreiserfsd
and here is the first few lines of top output from Box B
11:17pm up 3:00, 2 users, load average: 0.03, 0.19, 0.71
39 processes: 37 sleeping, 1 running, 1 zombie, 0 stopped
CPU states: 0.0% user, 3.7% system, 0.0% nice, 96.2% idle
Mem: 516996K av, 514392K used, 2604K free, 10892K shrd, 247208K buff
Swap: 136512K av, 96K used, 136416K free 234380K cached
PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND
900 niall 20 0 868 868 684 R 0 2.8 0.1 0:29 top
3 root 1 0 0 0 0 SW 0 0.9 0.0 11:59 kupdate
1 root 0 0 476 476 404 S 0 0.0 0.0 0:07 init
2 root 0 0 0 0 0 SW 0 0.0 0.0 3:34 kflushd
4 root 0 0 0 0 0 SW 0 0.0 0.0 0:00 kpiod
5 root 0 0 0 0 0 SW 0 0.0 0.0 0:44 kswapd
Notice the enormous amount of CPU chewed up by kupdate and kflushd on box B.
In the above snapshot, X is no longer running on box B. It was, and then I
started copying about 6GB of data from one disk to another. The response on
the machine slowed to a crawl. I was running a top, and noticed the huge
amount of time being used by kupdate and kflushd - kupdate regularly showing
>80 %CPU. Magicdev was also chewing up huge amounts of CPU for some reason
and there was no use whatsoever of floppy or CD. Eventually, with great
difficulty, I managed to control the mouse to log off Gnome. I then ran a
top in a console window, and that's what you see above. AFAIR kupdate was
still chewing up a huge chunk of the CPU until the copy was finished.
I just ran another copy on box B - about 700M of data copied by
find XXXX|cpio -pmd - and again kupdate and kswapd are taking up inordinate
amounts of CPU, while the cpio process was generally using upto 80%.
A similar copy on box B (less data, but same kind of load, I'd imagine i.e.
copying from IDE to the same IDE) shows the cpio process rarely taking up
10% of CPU and the interactive performance of the box is barely affected -
I'm writing this under X on that box, while box B with a couple of consoles
used (one with the copy, one with top, is as sluggish as anything). Any of
you bright sparks have any ideas ?
Regards,
Niall
Maintained by the ILUG website team. The aim of Linux.ie is to
support and help commercial and private users of Linux in Ireland. You can
display ILUG news in your own webpages, read backend
information to find out how. Networking services kindly provided by HEAnet, server kindly donated by
Dell. Linux is a trademark of Linus Torvalds,
used with permission. No penguins were harmed in the production or maintenance
of this highly praised website. Looking for the
Indian Linux Users' Group? Try here. If you've read all this and aren't a lawyer: you should be!
