Segfaults with Mapserver 5.0.0

[Rodrigo Del C. Andrade]

     Sure I can. In fact, got some results at hand.
     Attached is a "tail-ed" output from the last 150 lines of the log 
when it crashed, for comparison values. You get to see for the first 
page or so it processing the data during a zoom in, then you see a bit 
of java debug native to the application on the last few lines and when a
zoom out occurs all goes down.
     What I've found interesting in that last line just before the app
going tits up is the awfully high offset value. I don't have a clue what
that value stands for, but it seems out of place comparing with the
others. That layer in question is one being indexed by a tileindex, and
it was split from a larger shapefile using OpenJump.
     Also may be worth noting that I was having problems to get it to
crash with the patch. I was suspecting of a race condition being slowed
down by the extra work it had writing the ouput to file, so I asked a 
colleague to log on and give it at go simultaneously, and that did it. 
It went down in 30 secs with us both just grinding the zoom button (not 
before spewing 220 megs of logs :)

     I hope it helps.
     Log attached. Thanks a lot,

     Rodrigo

ps: re-sending with a smaller chunk of the log file. Previous one 
returned because the file was too big. Oops ;)


Umberto Nicoletti wrote:
> Rodrigo, could you apply the attached patch and the post the last
> output of the application before the segfault?
> Please note that it will log an awful LOT of stuff!
>
> Regards,
> Umberto
>
> On Jan 18, 2008 2:07 PM, Rodrigo Del C. Andrade
> <rodrigo.andrade at digitro.com.br> wrote:
>   
>>     Good morning all.
>>
>>     I am running tests on Mapserver 5 to test its viability on our
>> production enviroment and I have run into some problems I am having
>> trouble to figure out.
>>
>>     A little background, may be irrelevant, but could be useful, idk:
>>     - We have a stand alone java server application (doing the map work
>> with Java Mapscript) which talks via RMI with the client part of the
>> app, which is running on Tomcat. We've done it this way because past
>> crashes on mapserver used to bring Tomcat down with it
>>     - This mapscript binary was compiled with all the latest stable
>> versions of GDAL, GEOS, Proj, AGG, LibPNG, GD and Freetype.
>>     - We are outputting the final map images with AGG.
>>     - The source data lives in Shape files. All of them have indexes,
>> and I broke large areas into smaller chunks and tile indexed them with
>> tile4ms.
>>
>>     Now, this one is the first of the the 2 different segfaults  I am
>> experiencing with 5.0, but decided to focus on this one first because it
>> seems to happen more frequently. Whenever I am zooming in or out in the
>> map, this bugger seems to happen at random. I was'nt able to pin point a
>> pattern for it to happen, looks like just using it long enough will do it.
>>
>> #
>> # An unexpected error has been detected by HotSpot Virtual Machine:
>> #
>> #  SIGSEGV (0xb) at pc=0x00734600, pid=30922, tid=1812704176
>> #
>> # Java VM: Java HotSpot(TM) Server VM (1.5.0_07-b03 mixed mode)
>> # Problematic frame:
>> # C  [libc.so.6+0x69600]  strncpy+0x30
>> #
>>
>>     I am not any C buff by any means, but judging from a strncpy error,
>> I assume something in the way strings are created/handled might have
>> changed since 4.x (I haven't been following the list recently, forgive
>> me if it's  old news)
>>     Any ideas about what might be causing it?
>>
>>     -- Full log is bellow the sig.
>>
>>     Thanks!
>>
>> --
>>
>> *Rodrigo Del C. Andrade*
>> /Programador/
>> /SIC - SSE - Soluções Segurança Pública/
>>
>> *DÍGITRO TECNOLOGIA*
>> *E-mail:* rodrigo.andrade at digitro.com.br
>> <mailto:rodrigo.andrade at digitro.com.br>
>> *Fone:* +55 48 3281-7314 / +55 48 3281-7000
>> *Fax:* +55 48 3281-7299
>> *Site:* www.digitro.com <http://www.digitro.com

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: gis-server-debug-end.log
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/20080118/0ed2b5fa/attachment.ksh>