Re: [hobbit] Debugging issues with bb-histlog.sh on 4.3.0

[Alan Sparks <asparks (at) doublesparks.net>]

Ralph Mitchell wrote:
> On Wed, Jun 24, 2009 at 7:24 PM, Alan Sparks <asparks (at) doublesparks.net
> <mailto:asparks (at) doublesparks.net>> wrote:
>
>     T.J. Yang wrote:
>     > I am experiencing same problem but it is on a  4.2.0 server.
>     > If you like, we can work together to debug this issue.
>     > This bug is listed on my hobbit deployment todo list .
>     >
>     > --------------------------------------------------
>     > From: "Alan Sparks" <asparks (at) doublesparks.net
>     <mailto:asparks (at) doublesparks.net>>
>     >> I've some "random" failures on bb-histlog.sh, viewing items in
>     the bb2
>     >> event log.  Some work as expected, but some when clicked appear
>     to do
>     >> nothing.  Hitting the URL for that (say
>     >>
>     http://den2sch17/hobbit-cgi/bb-histlog.sh?HOST=den1ws063&SERVICE=nfs&TIMEBUF=Mon_Jun_22_18:46:39_2009
>     <http://den2sch17/hobbit-cgi/bb-histlog.sh?HOST=den1ws063&SERVICE=nfs&TIMEBUF=Mon_Jun_22_18:46:39_2009>)
>     >>
>     >> with curl shows a zero-length reply.  Checking the Apache error log
>     >> shows "Premature end of script headers: bb-histlog.sh, referer:
>     >> http://den2sch17/hobbit/bb2.html";.  I have found no core files left
>     >> behind.
>     >>
>     Well, I have mixed luck running under GDB... looking at the code,
>     saw it
>     was trying to get the CGI info from the environment so set some
>     environment variables to fake up enough CGI environment to get it to
>     run.  But mostly wound up getting segfaults in expand_env on the
>     HTMLCONTENTTYPE stuff.
>
>     But I stumbled upon something that fixed my problem -- but not sure
>     why.  I edited the cgi-bin/bb-histlog.sh file, and added
>     "cd /tmp" before the meat of the script.  For some reason, I now
>     seem to
>     have no issues.
>
>
> Perhaps that ought to be "cd $BBTMP" instead??  Keep all Xymon stuff
> in one place.
>
> Ralph Mitchell
>  
BBTMP is not defined when this script runs...
-Alan