Maybe it's not insufficent memory

[Lincoln Peters]

I think that I've eliminated insufficent memory as a suspect:
I noticed that Ximian's "Red Carpet" was also causing the computer to lock 
up.  So on Monday, I logged into FVWM2 and opened 2 XTerms.  In one I 
entered the following command:

while [ true ] ; do uptime ; free ; /etc/init.d/xfs status ; sleep 10s ; 
done

In the other one, I ran red-carpet.  As Red Carpet started up, I watched the 
other XTerm to see how Red Carpet affected the memory and if (by some 
miracle) I could see a clear sign if and when xfs dropped dead.  The same 
crash happened, but the memory didn't come close to running out (including 
buffers and caches that could have easily been flushed, I was reading a good 
10MB available before the crash).  Then, since it didn't respond to 
CTRL-ALT-Backspace or CTRL-ALT-DEL, I pressed the 'Reset' button.

So I rebooted and tried again, this time monitoring the memory and xfs 
status from tty2 while FVWM2 and red-carpet ran on tty7 (in the hope that I 
could narrow it down to an xfs or non-xfs problem).  As red-carpet started 
up, I kept the screen on tty2, and all of a sudden, it stopped giving me the 
read-outs and it didn't respond to CTRL-C.  Then I switched to tty7, and got 
a blank screen (a.k.a. complete X freeze).  Once again, I had to do a hard 
reboot.

I would have logged in from my server using SSH so that I could watch what 
was happening on two displays, but for some reason sshd on the workstation 
doesn't work anymore.  When I tried to use it, I got the error:

ssh_exchange_identification: Connection closed by remote host

And, no, I did NOT try again with rlogin or telnet.


Today (Tuesday), I tried to find to find any other applications that would 
cause the same crash.  I started AbiWord and typed up a memo that explains 
what my project is (code-named Firewalker) and what it can do. It took about 
30 seconds to load AbiWord, but it worked just fine.  But as soon as I 
finished typing up the memo, I was called away.  At least I was able to log 
out quickly and painlessly.

What I really should do is find a way to boot from one of the 3D animation 
computers.  They are dual-processor Pentium-III 733MHz's, and have 256MB of 
RAM, so if I can find out what kind of NIC's they have, I can boot one of 
them off of my server and probably determine with accuracy if these crashes 
are indeed connected to the small memory of the common workstations.


I just thought that some of you would like to know how this is going.  I 
still haven't found any common thread that links all of the different 
programs that lock up the workstation (and the ramdisk mounted at /var/log 
isn't helping).

_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp