[cobalt-users] RaQ2 freezing after issuing chown and cp commands, not first time happened

["Steven Werby" <steven-lists@xxxxxxxxxxxx>]

I was just on a client's RaQ2 installing majordomo (client needs a second,
modified version).  After issuing the following command from telnet the server
hung.

chown -R root:root *

I was working within a directory I created called /home/befriend/majordomo and
here is the exact history of commands from the telnet session to that point:

su -
cd /home/befriend/majordomo (temporary work directory)
tar xfvz majordomo-1.94.5.tar.gz (file that I copied into this directory
previously)
chown -R root:root *

After the telnet session hung I opened a second telnet session and ran "top".
Load average was under 0.5.

17347 root       0   0   524  524   440 D       0  0.0  0.3   0:00 chown

Notice that the process has a "D" status.  I am unable to kill it, yet it's
causing problems.  Load average continued to grow even though the server is
lightly loaded.  The "chown" command was successful though since the directory
files are owned by root:root.

I have encountered this problem twice before on a RaQ2 and twice on a RaQ3.  I'm
using SecureCRT by VanDyke as a telnet client.  On all occasions I was executing
the command "cp" or "chown".  Load average went from under 1.0 to over 10.0 and
I tried accessing the GUI and the browser hung while "loading page".  Every time
I executed "/sbin/shutdown -r" to shutdown the server and reboot it, but both
times the server did not reboot and the server had to be physically rebooted.
The load on the server has currently stabilized at:

3:24am  up 91 days, 11:48,  3 users,  load average: 5.00, 4.97, 4.22

I tried accessing the GUI and it hangs while "loading page" as it did the
previous times I encountered this problem.  I continued with the install in the
2nd telnet session and I was able to use "cd" to change directories and "pico"
to edit a file, but when I tried to do a simple one file copy using "cp -p" it
hung again!  It's late here and the client's in Greece so I'm not going to
execute the shutdown command this time.

Has anyone encountered this on the RaQs?  Is this a known bug with a patch in a
Cobalt package?  I'm pretty active on the Cobalt User list, but I don't recall
this issue coming up before.  Is there anything that can be done short of a
physical reboot?  In the past I've also tried restarting all relevant services
without any luck.

--
Steven Werby {steven-lists@xxxxxxxxxxxx}