Segmentation Fault - but Shibboleth SP works fine

[David Cardozo]

I have two installations of Shibboleth SP v2.5.2 in two different Solaris 10
servers; they both exhibit the following behavior: When I load the mod_shib
module in Apache and bounce Apache, a "Segmentation Fault - core dumped"
message is displayed in the console and a core file is created; however,
Apache does start and so far I haven't experience any other issues: I can
access my applications, Shibboleth behaves as expected, etc. In other words,
other than the Segmentation fault, nothing else seems wrong!!

I use the dbx tool to look inside the core file and here is the output:

-bash-3.00# dbx - /usr/local/apache2.2.17_test/bin/core
Corefile specified executable: "/usr/local/apache2.2.17_test/bin/httpd"
For information about new features see `help changes'
To remove this message, put `dbxenv suppress_startup_message 7.7' in your
.dbxrc
Reading httpd
core file header read successfully
Reading ld.so.1
Reading libz.so.1
Reading libssl.so.1.0.0
Reading libcrypto.so.1.0.0
Reading libdl.so.1
Reading libm.so.2
Reading libaprutil-1.so.0.3.10
Reading libexpat.so.0.5.0
Reading libapr-1.so.0.4.2
Reading libuuid.so.1
Reading libsendfile.so.1
Reading librt.so.1
Reading libsocket.so.1
Reading libnsl.so.1
Reading libpthread.so.1
Reading libc.so.1
Reading libaio.so.1
Reading libmd.so.1
Reading libc_psr.so.1
Reading libscf.so.1
Reading libdoor.so.1
Reading libuutil.so.1
Reading libgen.so.1
Reading libmd_psr.so.1
Reading libmp.so.2
Reading mod_shib_22.so
Reading libshibsp-lite.so.6.0.2
Reading libxmltooling-lite.so.6.0.3
Reading libxerces-c-3.1.so
Reading libicuuc.so.3
Reading libicudata.so.3
Reading liblog4shib.so.1.0.8
Reading libstlport.so.1
Reading libCrun.so.1
t at 1 (l at 1) program terminated by signal SEGV (no mapping at the fault
address)
0xfe69020c: ~locale+0x0008:     ld       [%o3 + 16], %o2
Current function is destroy_and_exit_process
  272       exit(process_exit_value);
(dbx) where
current thread: t at 1
  [1] std::locale::~locale(0xfe003c58, 0x2b4, 0xfdfff374, 0x0, 0xfeff2a00,
0x1c00), at 0xfe69020c
  [2] _exithandle(0xfedb7500, 0xfedb5900, 0x1c00, 0x0, 0xffbffb50, 0x0), at
0xfecc27d8
  [3] exit(0x0, 0x2710, 0x756c00, 0x756c0048, 0x80808080, 0x1010101), at
0xfecb07d0
=>[4] destroy_and_exit_process(process = 0x12b888, process_exit_value = 0),
line 272 in "main.c"
  [5] main(argc = 3, argv = 0xffbffcdc), line 662 in "main.c"

>From the above, looks like an issue in Apache when destroying child process;
but again, only happens when mod_shib is loaded. Apache has been compiled
with the same version of C in sunstudio12.3 as Shibboleth.

My questions are:
* Have anyone else experienced this problem? (I've seen other "Segmentation
fault" issues in the mailing list, but none seems to describe this "fails,
but works anyway" scenario)
* Are there any suggestions I should try? (other than moving away from
Solaris ­ this will happen eventually)
Thanks,

David.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://shibboleth.net/pipermail/users/attachments/20131113/c5df9696/attachment.html