[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: server debug



ahh, i had a vague thought that might be the case. The confusing thing is that 
you were also hitting the 9080 server, so i was seeing some of your accesses 
there.

yes, we keep the "development server" up most of the time, but its less stable.
likely i will move it to the production server (8080) next week, i will post to 
thredds email group when that happens.

thanks for your help in debugging...


David Wojtowicz wrote:
Actually.... nevermind.

I just noticed that in copying files about, I had gone back to using my
original config file that was accessing the 8080 server, which is presumably
still the old code with bugs included.
When using the 9080 server, I am no longer getting any problems.   So, I
think you've solved the problems!

If you wouldn't mind keeping the 9080 server going until you move your fixed
code the production server, I'd appreciate it.

Thanks for all your help.


-----
David Wojtowicz, Sr. Research Programmer, Sysadmin
Dept of Atmospheric Sciences / Computer Services
University of Illinois at Urbana-Champaign
address@hidden  (217) 333-8390

-----Original Message-----
From: David Wojtowicz [mailto:address@hidden] Sent: Thursday, March 09, 2006 2:14 PM
To: 'John Caron'
Subject: RE: server debug

Hi John,

 I'm still having problems.   Sometimes, it does the "Bad Request" (same
thing as the packet trace I sent you earlier)  Othertimes, it hangs during
the initial connect.  Othertimes, it does a few variables and then just
seems to sit there (If I break it in the debugger, it is stopped in the
select() call inside Curl... apparently waiting on the server).
Othertimes it works fine. (sometimes many times in a row)

I've distilled out a chunk of my code as a separate test program that you
can try.   (Maybe its just me.... though it happens on both my linux boxes
and Mac OSX box).

The code is here:

 http://www.atmos.uiuc.edu/~davidw/testprog.tar.gz

You'll need to modify the variable at the top of Makefile to point to where
DODS is.   It requires my string patch on libnc-dap.   If you want to use
straight DODS without my string patch, then you need to comment out line 64
of testprog.cc:

 char *stnids=n.getCharSlab("station_id",ic);

which will fail without the patch.


The script "test-driver" can be used to loop it over and over.


-----
David Wojtowicz, Sr. Research Programmer, Sysadmin
Dept of Atmospheric Sciences / Computer Services
University of Illinois at Urbana-Champaign
address@hidden  (217) 333-8390

-----Original Message-----
From: John Caron [mailto:address@hidden] Sent: Wednesday, March 08, 2006 8:10 PM
To: David Wojtowicz
Subject: Re: server debug

Hi David:

Im pounding away on the "latest" metar file, without being able to generate
that bug. So there is something in your software that i havent yet
reproduced. Also, I see your requests in my logs, but no error is being
generated.

If you get a chance to try it again, and can capture as much info as
possible about what exactly is happening, i would be most grateful.

thanks,
john