How much do I need to worry about this?

Richard Kreiss rkreiss at gccconsulting.net
Tue May 12 19:04:41 PDT 2009



> -----Original Message-----
> From: filepro-list-bounces+gccconsulting=comcast.net at lists.celestial.com
>
[mailto:filepro-list-bounces+gccconsulting=comcast.net at lists.celestial.com]
On
> Behalf Of Boaz Bezborodko
> Sent: Tuesday, May 12, 2009 3:33 PM
> To: Nancy Palmquist
> Cc: filepro-list at lists.celestial.com
> Subject: Re: How much do I need to worry about this?
> 
> 
> 
> Nancy Palmquist wrote:
> > Boaz Bezborodko wrote:
> >> John Esak wrote:
> >>
> >>>  Well, I guess you mean a comment on the if: line of a lookup.  I
don't
> >>> think anything other than the lookup can go on a lookup line.  But,
> >>> still,
> >>> shouldn't cause rcabe to crash.  It's sounding like an overall
> >>> memory error.
> >>> I didn't understand what you were saying about your testing on other
> >>> machines.  Did it not crash there so you think it is just on your one
> >>> development machine?  Is it at all repeatable?  Have you eliminated
> >>> hardware
> >>> problems.  Re-seated the memory, simple things like that?
> >>>
> >> The problem happens when I have a syntax error in the table and
> >> occurs no matter which machine I use in the office.  I checked the
> >> log on Samba and it seems to be opening and closing the files as
> >> necessary and isn't complaining about anything.
> >> I also tried copying all the files from the file server onto my own
> >> machine and set PFDSK to the local disk and I still got the problem.
> >> This should have eliminated any memory problems (wouldn't duplicate
> >> on other machines) and the server (I'm running the programs off the
> >> local disk).
> >>
> >>
> >>> I do know one error that causes very strange behavior with cabe... I
> >>> don't
> >>> rmember if it does actual crashes. It might... Though... I know it
> >>> does all
> >>> sorts of weird stuff.  This is the error of not having enough memory
> >>> allocated to "label" size. There is an environment variable that can
> >>> be set
> >>> higher.  PFLBSIZE.  As you get more and more lookup aliases, array
> >>> names I
> >>> think, and actual labels, maybe some other things, Ken would have to
> >>> elaborate, the tables can not be saved properly and lots of strange
> >>> runtime
> >>> errors occur.  I set it in most of my large systems to 5000.  I
> >>> think it is
> >>> defaulted to only 1024.  Like I say, it probably doesn't cause cabe to
> >>> crash, but then again, that might be one of the symptoms.
> >>>
> >>> John
> >>>
> >>>
> >>>
> >> This couldn't be the problem as I can make it do this even if the
> >> only line in the table is a lookup with a syntax error.
> >>
> >> One thing more....I'm seeing the following on the screen right after
> >> the error window pops up:
> >>
> >> Exception error -1073741819 in process failed 0(0x00000000)
> >> comspec="(null)"
> >> command="\fpro\fp\rcabe
> >>          "
> >> Press Enter:
> >>
> >>
> >> _______________________________________________
> >> Filepro-list mailing list
> >> Filepro-list at lists.celestial.com
> >> http://mailman.celestial.com/mailman/listinfo/filepro-list
> >>
> >>
> > Are you running a WINDOWS version of filePro?  I had the impressing
> > you were running a Unix/Linux version from the talk of Samba.
> >
> > I had no end of issues with filePro when the Windows network is not
> > working correctly.
> >
> > The errors were like SegViolations on Unix and reported like you see
> > above.  It always seemed to be caused by the system being unable to
> > write fast enough.
> >
> > Now mostly I saw this in the runtime side, but maybe your network has
> > other problems.
> >
> > To fix my RUNTIME issues, I added SLEEP commands after WRITE or
> CLOSE
> > to give the network time to write the data and update the server.
> >
> > It fixed most of the problems.  It has reduced them to the point where
> > they are very  occasional, instead of crashing so much we could not
work.
> >
> > Check that you have applied all the network patches for the system,
> > especially related to LOCKING and I/O, for the server and clients.
> >
> > Hope that will help.
> >
> > Nancy
> Thanks for the in depth responses, Nancy.
> 
> I am using FilePro running locally on Windows machines and using a Linux
> server running Samba as the fileserver.  (When we made this change we
> were already running FP Windows off a Netware server, so moving to Samba
> when we needed to upgrade seemed the logical decision.)
> 
> I have no problems with rclerk or rreport.  In fact everyone else is
> moving along fine without even knowing about what is happening (except
> for my having to explain why I can't get anything else done right now).
> The error occurs as soon as I enter 'Y' at Check Syntax.  Yes, it saves
> the file, but it won't compile it.
> 
> To isolate the problem I moved everything over to my own local machine
> and change PFDSK to point to the local drive.  It still happens which
> means that it isn't related to the server.  I also get the same problem
> on other machines so it can't be a memory issue.
> 
> I ran Microsoft's FILEMON to see what the system was doing with the
> files and I did find one interesting thing when I looked at the data.  I
> found that rcabe was looking for 3 files that I can't find on my system.
> 
> 273    2:53:40 PM    rcabe.exe:360    QUERY INFORMATION
> E:\fpro\fp\WINSTA.dll    NOT FOUND    Attributes: Error
> 274    2:53:40 PM    rcabe.exe:360    QUERY INFORMATION
> E:\fpro\fp\WTSAPI32.dll    NOT FOUND    Attributes: Error
> 275    2:53:40 PM    rcabe.exe:360    QUERY INFORMATION
> E:\fpro\fp\SETUPAPI.dll    NOT FOUND    Attributes: Error
> 
> I have sent FPsupport this information as well as the complete FILEMON
> data from the error.
> 
If you haven't already done so, set abe=ASCII and then resave the processing
table without the syntax check.  Use an editor or "type foobar |more" to
check  to see if there are any odd characters in the processing table.

There may be some graphics characters hiding.


Richard Kreiss
GCC Consulting
rkreiss at gccconsulting.net
  






More information about the Filepro-list mailing list