[ZOO-Discuss] Execute request returns SIGSEGV message from the zoo kernel for Python

Tue Mar 22 10:06:51 PDT 2011

Gerald

I used the command you mentioned:

./zoo_loader.cgi
"request=Execute&service=WPS&version=1.0.0&Identifier=HelloPy&DataInputs=a=Demo"

curl version is 7.15.5

I compiled again with -DTIGER and the command still causes a segfault,
backtrace is then:

Core was generated by `./zoo_loader.cgi
request=Execute&service=WPS&version=1.0.0&Identifier=HelloPy&D'.
Program terminated with signal 11, Segmentation fault.
#0  0xb74194db in PyObject_Malloc () from /usr/lib/libpython2.5.so.1.0
(gdb) bt
#0  0xb74194db in PyObject_Malloc () from /usr/lib/libpython2.5.so.1.0
#1  0xb74224da in PyString_FromString () from /usr/lib/libpython2.5.so.1.0
#2  0xb7413d73 in PyDict_SetItemString () from /usr/lib/libpython2.5.so.1.0
#3  0xb7478e66 in PyImport_Cleanup () from /usr/lib/libpython2.5.so.1.0
#4  0xb7484d91 in Py_Finalize () from /usr/lib/libpython2.5.so.1.0
#5  0x080593ec in zoo_python_support (main_conf=0xbf972848,
request=0x806e988, s=0x8073028, real_inputs=0xbf972840,
real_outputs=0xbf972844)
    at service_internal_python.c:195
#6  0x0804c23b in loadServiceAndRun (myMap=0xbf982cfc, s1=0x8073028,
request_inputs=0x806e988, inputs=0xbf982cf0, ioutputs=0xbf982cf8,
eres=0xbf982ce8)
    at zoo_service_loader.c:306
#7  0x0804ee5d in runRequest (request_inputs=0x806e988) at
zoo_service_loader.c:1770
#8  0x0804b201 in cgiMain () at zoo_loader.c:264
#9  0x080603dc in main (argc=2, argv=0xbf982e84) at cgic.c:303

I will checl on a nother machine as well.
armin

On Tue, Mar 22, 2011 at 5:57 PM, Gérald Fenoy <gerald.fenoy at geolabs.fr> wrote:
> Armin,
> it seems that the issue is lightly different from the first one.
> Maybe it comes from the cURL version you are using. To confirm that
> assumption, can you
> try to add -DTIGER to the CFLAGS in your Makefile to then clean and
> recompile the
> ZOO-Kernel ?
> If it works this mean that we have to look deeply in the cURL version
> installed on the target
> and check which one cause trouble and which one not. Or force installer to
> use a specific cURL
> version.
> I'm currently use ZOO-Kernels linked against 7.19.6 (on GNU/Linux)
> and 7.19.7 (on MacOS X)
> without issue.
> Something sounds odd to me in your last backtrace. Indeed it seems that the
> ZOO-Kernel
> is trying to download something but in your example it should not. Can you
> give me the exact
> command line you were using for getting such a result ?
> Hope to hear from you,
> Best,
> Le 22 mars 2011 à 17:29, Armin Burger a écrit :
>
> Gerald
>
> I followed all the steps and still get the segfault message
>
> <ows:Exception exceptionCode="InternalError">
>    <ows:ExceptionText>ZOO Kernel failed to process your request
> receiving signal 11 = SIGSEGV</ows:ExceptionText>
>  </ows:Exception>
>
> If I compile with USE_DBG the it writes out a core dump file,
> analyzing it gives the message
>
> Core was generated by `./zoo_loader.cgi
> request=Execute&service=WPS&version=1.0.0&Identifier=HelloPy&D'.
> Program terminated with signal 11, Segmentation fault.
> #0  0xb7f317ea in curl_formadd () from /usr/lib/libcurl.so.3
> (gdb) bt
> #0  0xb7f317ea in curl_formadd () from /usr/lib/libcurl.so.3
> #1  0xb7f41018 in curl_slist_free_all () from /usr/lib/libcurl.so.3
> #2  0xb7f4cf67 in curl_easy_setopt () from /usr/lib/libcurl.so.3
> #3  0x08059aaf in InternetOpen ()
> #4  0x0804d4f0 in runRequest (request_inputs=0x806e988) at
> zoo_service_loader.c:708
> #5  0x0804b201 in cgiMain () at zoo_loader.c:264
> #6  0x08060414 in main (argc=2, argv=0xbf88d584) at cgic.c:303
>
>
> armin
>
> On Tue, Mar 22, 2011 at 4:39 PM, Gérald Fenoy <gerald.fenoy at geolabs.fr>
> wrote:
>
> Hi Armin,
>
> I did the following steps and everything goes right.
>
> 1) Download the SVN source tree
>
> 2) compile cgic206 from the SVN source tree thirds  directory
>
> 3) go into zoo-kernel directory and run autoconf
>
> 4) run : ./configure --with-python
>
> 5) create a directory zoo-dist
>
> 6) copy zoo_loader.cgi and main.cfg in zoo-dist
>
> 7) copy zoo-services/hello-py/test_service.py and
>
> zoo-services/hello-py/cgi-env/HelloPy.zcfg in zoo-dist
>
> 8) from the zoo-dist directory, I run the following :
>
> ./zoo_loader.cgi
>
> "request=Execute&service=WPS&version=1.0.0&Identifier=HelloPy&DataInputs=a=Demo"
>
> I get the hello world message included in the ExecuteResponse.
>
> Please can you try to reproduce the steps above on your platform and
>
> let me know if the error still occurs ?
>
> Hope to hear from you,
>
> Best
>
> Le 22 mars 2011 à 14:16, Armin Burger a écrit :
>
> Nick
>
> I tried the SVN from today, still a segfault. I tried using GDB and
>
> the only debug info I get is:
>
> Core was generated by `./zoo_loader.cgi
>
> Service=WPS&Request=Execute&Version=1.0.0&Identifier=HelloPy&D'.
>
> Program terminated with signal 11, Segmentation fault.
>
> #0  0xb70c4a61 in fileno_unlocked () from /lib/tls/libc.so.6
>
> (gdb) bt
>
> #0  0xb70c4a61 in fileno_unlocked () from /lib/tls/libc.so.6
>
> #1  0xb73ef188 in FCGI_fileno () from /usr/lib/libfcgi.so.0
>
> #2  0x0804b039 in cgiMain () at zoo_loader.c:97
>
> #3  0x080602b3 in main (argc=2, argv=0xbfd81284) at cgic.c:287
>
>
> armin
>
>
> On Thu, Mar 17, 2011 at 3:42 PM, nicolas bozon <nicolas.bozon at gmail.com>
>
> wrote:
>
> Hi Armin,
>
> You are right, it seems we get a problem, as we could reproduce the segfault
>
> this morning before going to the code sprint.
>
> We'll try to solve the problem with gerald today and we'lll keep you
>
> informed.
>
> Best,
>
> Nick
>
> 2011/3/14 Armin Burger <armin.burger at gmail.com>
>
> Hi all
>
> after successfully compilation and installation of the zoo_loader.cgi
>
> the "GetCapabilities" and "Describe Process" work fine. But as soon as
>
> I try to run the Excecute request  using the simple HelloPy and a
>
> python serviceProvider as hello.py with
>
>
> http://my.website/zoo/?Service=WPS&Request=Execute&Version=1.0.0&Identifier=HelloPy&DataInputs=a=My%20Name
>
> the zoo kernel returns a segmentation fault message like
>
> <?xml version="1.0" encoding="UTF-8"?>
>
> <ows:ExceptionReport xmlns:ows="http://www.opengis.net/ows/1.1"
>
> xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
>
> xmlns:xlink="http://www.w                  3.org/1999/xlink"
>
> xsi:schemaLocation="http://www.opengis.net/ows/1.1
>
> http://schemas.opengis.net/ows/1.1.0/owsExceptionReport.xsd"
>
> xml:lang="en-                  US" version="1.1.0">
>
>  <ows:Exception exceptionCode="InternalError">
>
>    <ows:ExceptionText>ZOO Kernel failed to process your request
>
> receiving signal 11 = SIGSEGV</ows:ExceptionText>
>
>  </ows:Exception>
>
> </ows:ExceptionReport>
>
> Same happens when I run it from the command line
>
> ./zoo_loader.cgi
>
> "Service=WPS&Request=Execute&Version=1.0.0&Identifier=HelloPy&DataInputs=a=My%20Name"
>
> So it seems sort of "caught" segfault from some internal
>
> functionality. But I do not know how to do some kind of debugging.
>
> Any ideas what I could try?
>
> Armin
>
> _______________________________________________
>
> Zoo-discuss mailing list
>
> Zoo-discuss at gisws.media.osaka-cu.ac.jp
>
> http://gisws.media.osaka-cu.ac.jp/mailman/listinfo/zoo-discuss
>
>
> _______________________________________________
>
> Zoo-discuss mailing list
>
> Zoo-discuss at gisws.media.osaka-cu.ac.jp
>
> http://gisws.media.osaka-cu.ac.jp/mailman/listinfo/zoo-discuss
>
> Gérald Fenoy
>
> gerald.fenoy at geolabs.fr
>
>
> Gérald Fenoy
> gerald.fenoy at geolabs.fr
>