on both 2.1.127 and 2.1.128, heavy exec() (with a fair bit of I/O)
seems to cause the system to go into a busy wait state. during this
time, it is difficult/impossible/very slow to launch any new
processes.
i have a small c program which can usually trigger this behavior,
although it's somewhat erratic. this c program just forks, execs
something, waits for it, and does it again. (but read the comments as
the top, as i kept on adding different things to try...)
http://www.scrye.com/~tkil/fork-test.c
if the "something" is trivial (either just exit(), or call something
like "wc -c"), then it just screams along. if the "something" is more
complex (i originally ran into this bug when trying to use djpeg to
verify 17000 jpeg images totalling some 1.2GB), then it usually will
reach a point where it looks like it can't start any new processes.
i'm currently testing with:
fork-test 3 /usr/bin/djpeg -fast -outfile /dev/null < pics.txt
the rest of the system is responsive, and there is no heavy disk i/o
or other signs of a locked-up box. and it looks like this problem
eventually goes away by itself; i'm running my test program for the
second time (looks like first time was stuck for a while, but it
finished eventually).
stats:
AMD K6-2 300 @ 100MHz bus, 128MB RAM
FIC PA-2013A mainboard (1MB cache, VIA-MVP3 chipset)
Asus SC875 (NCR53C875) SCSI HA (using ncr53c8xx driver:
CONFIG_SCSI_NCR53C8XX=y
CONFIG_SCSI_NCR53C8XX_DEFAULT_TAGS=8
CONFIG_SCSI_NCR53C8XX_MAX_TAGS=32
CONFIG_SCSI_NCR53C8XX_SYNC=20
RedHat 5.1 base, various add-on packages.
compiler:
Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/egcs-2.90.27/specs
gcc version egcs-2.90.27 980315 (egcs-1.0.2 release)
glibc2: 2.0.7
if anyone would like further information, please just drop me a line
and i'd be glad to get it for you.
thanks,
t.
--
Tkil * <URL: http://www.scrye.com/~tkil> * hopelessly hopeless romantic.
"So amplify this little one | She hears as much as she can see
She's a volume freak | And what she sees, she can't believe."
-- Catherine Wheel, _Happy Days_, "Judy Staring At The Sun"
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/