NFS stops responding

[Kevin D. Clark]

Derek Atkins writes:

> Will you share the results once your analysis is done?

It's hard for me to go into many specifics, but maybe this will
satisfy your request: the network I was dealing with was a fairly
high-security network (really...) and through some investigation I
learned that one of the {things} that helped protect this network was
a transparent proxy...and I learned that some users of this
transparent proxy reported that certain versions of this proxy
interfered with certain types of CORBA/GIOP traffic.  Well, at this
point I *had to* try to rule out this scenerio, so I performed my
network analysis.  After quite a bit of analysis, I concluded that the
proxy was not at fault.  Indeed, the network seemed to be working
properly.

So, at this point, I was again at square one, but at least I reduced
the size of my problem domain.

Even at this point, the traces that I captured were useful.  I went
back through them again after my initial analysis and this time asked
myself the question "OK, things are starting to go wrong right *here*
-- now, where in the code are we executing when things start to go
bad?".  After some more investigation, I found a subtle class of
coding...foible near the {thing} that I was trying to debug.  So, I've
been fixing this foible in the {thing} -- all the while being unable
to reproduce the overall problem in a lab.  However, as luck would
have it, somebody I know reproduced the problem today (it is very hard
to reproduce), and I was able to get a good sense that the fix that I
am implementing will indeed address this problem.

I'm going to feel so good when I checkin and get this problem off my
back!  I have some more unit tests to run now...

Regards,

--kevin
-- 
alumni.unh.edu!kdc / http://kdc-blog.blogspot.com/
GnuPG: D87F DAD6 0291 289C EB1E 781C 9BF8 A7D8 B280 F24E

 Wipe him down with gasoline 'til his arms are hard and mean
 From now on boys this iron boat's your home
 So heave away, boys.
   -- Tom Waits