Re: Random processes hanging in unkillable state in -BETA6

 [ I don't know / think is related, just gathering info ]

On Wed, 29 Sep 2004 00:57:23 -0700
Julian Elischer <julian_at_elischer.org> wrote:

> Damian Gerow wrote:
> > Thus spake Julian Elischer (julian_at_elischer.org) [29/09/04 00:48]:
> > : If you look at them with ps -alxH are they all (in particular dig)
> > : all threaded programs?
> > 
> > Yes, they are.
> > 
> > They all seem to be in an RL state.  All the dig processes are in RL+,
> > whereas dig processes started by root are in RXL+.  There is one hung gdb
> > dig process, which is in SL.  Note that these have all been running for at
> > least three hours at this point.
> 
> the time is not important they will never recover..
> you might do the following.. if you have ddb in the kernel..
> 
> go to console
> <CTL><ALT><ESC> to go to ddb


Will the system recover if I have X running (my previous experiences
says no, but I haven't break in debugger for some time now)?

I have a dcgui-qt hang when I put it to recreate the share list (this is
after I've tried kill -1,-6,-9) :

 # ps -alxH 75295
  UID   PID  PPID CPU PRI NI   VSZ  RSS MWCHAN STAT  TT       TIME COMMAND
 1001 75295 88546   0  76 -10 105252 94608 -      T<XL  ??  200:59.34 /usr/local/bin/dcgui-qt
 1001 75295 88546   0  76 -10 105252 94608 -      T<XL  ??  200:59.34 /usr/local/bin/dcgui-qt
 1001 75295 88546   0  20 -10 105252 94608 -      T<XL  ??  200:59.34 /usr/local/bin/dcgui-qt
 1001 75295 88546   0   8 -10 105252 94608 wait   T<XL  ??  200:59.34 /usr/local/bin/dcgui-qt
 1001 75295 88546   0  20 -10 105252 94608 -      T<XL  ??  200:59.34 /usr/local/bin/dcgui-qt

 # uname -a
FreeBSD it.buh.tecnik93.com 5.3-BETA5 FreeBSD 5.3-BETA5 #0: Fri Sep 24 05:51:02 EEST 2004     itetcu_at_it.buh.tecnik93.com:/usr/obj/usr/src/sys/IT53_d  i386

 # sysctl debug.witness
debug.witness.skipspin: 1
debug.witness.trace: 1
debug.witness.kdb: 0
debug.witness.watch: 0

> ps
> look for the process.. note that several threads show up..
> there will be a kg (ksegrp) associated with ech thread (some
> may be associated with > 1 thread.)
> 
> of the address of the kg is 0xAB123456
> then do:
> x/xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx 0xAB123456
> 
> for each kg
> 
> and let me see the result(s)..

How much text is it supposed to be (I don't have a serial console on
this machine).


Thanks,

-- 
IOnut
Unregistered ;) FreeBSD "user"