Problems with uthread_kern.c in FreeBSD-5.1-p10

[Kris von Mach]

Hello,

I am not sure if this is the right place to post this, but here it goes.

There seems to be problems with uthread_kern.c and maybe other parts of 
libc/threads in FreeBSD-5.1-p10.

Here are some details about the system:

FreeBSD vorlon.swishmail.com 5.1-RELEASE-p10 FreeBSD 5.1-RELEASE-p10 #0: 
Mon Oct  6 02:16:15 EDT 2003 
root at vorlon.swishmail.com:/usr/obj/usr/src/sys/VORLON  i386

CPU: AMD Opteron(tm) Processor 246 (1990.05-MHz 686-class CPU)
real memory  = 3690987520 (3520 MB)
avail memory = 3590627328 (3424 MB)
FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs

I have installed Apache2, mod_php4, openssl from ports.

Apache/2.0.48 (Unix) PHP/4.3.4 mod_ssl/2.0.48 OpenSSL/0.9.7c

Apache starts up fine. But after a while httpd process ends up using 100% 
CPU, and eventually more httpd processes start doing the same and the web 
server becomes unresponsive.

  PID USERNAME   PRI NICE   SIZE    RES STATE  C   TIME   WCPU    CPU 
COMMAND
59780 nobody     130    0 36080K 20316K *Giant 1  19:57 94.09% 94.09% httpd

Nov 14 16:22:55 vorlon kernel: pid 59390 (httpd), uid 65534: exited on 
signal 6
Nov 14 16:25:33 vorlon kernel: pid 59393 (httpd), uid 65534: exited on 
signal 6
Nov 14 16:28:10 vorlon kernel: pid 59672 (httpd), uid 65534: exited on 
signal 6

I have been trying to track what is causing this and enabled debug in 
httpd.conf and I started getting these errors in error_log for apache:

Fatal error 'Unable to read from thread kernel pipe' at line 1100 in file 
/usr/src/lib/libc_r/uthread/uthread_kern.c (errno = 0)
[Fri Nov 14 16:28:10 2003] [notice] child pid 59672 exit signal Abort trap 
(6)

I believe someone posted that Gnucash, and ruby have problems and give the 
same error as above. Also there was a post about Apache2 and SSL locking up 
the system, and this may be related.

On another note, I have installed Setiathome, and when ran it only starts 
with one process instead of two like it should:

/usr/local/etc/rc.d/setiathome.sh start
Another processes is already using the shared memory.
 SETI at home

ps -auxw |grep seti
setiathome 62458 99.0  0.4 16232 15228  p2  RN    5:37PM   5:38.64 
/usr/local/sbin/setiathome -email -graphics -nice 15
setiathome 62457  0.0  0.0   900  484  p2  I     5:37PM   0:00.00 /bin/sh 
-T /usr/local/libexec/setiathome.bin
setiathome 62465  0.0  0.0   144   36  p2  I     5:37PM   0:00.00 sleep 
21600

I have not experienced any of these problems on FreeBSD-4.8 or earlier. 
Though I haven't tried running apache2 on them yet.

Is this a bug or a user problem?