Without apache-mod_perl-2.0.5-4.mga2 installed: # service -f httpd Starting httpd: [ OK ] With mod_perl installed: # service -f httpd Starting httpd: ./httpd: line 165: 5800 Segmentation fault $httpd $httpdconf $extramodules $modules $options (wd: /etc/init.d) [ÃCHEC ]
Hi Frédéric, (In reply to comment #0) > Without apache-mod_perl-2.0.5-4.mga2 installed: > > # service -f httpd > Starting httpd: [ OK ] > > With mod_perl installed: > > # service -f httpd > Starting httpd: ./httpd: line 165: 5800 Segmentation fault $httpd > $httpdconf $extramodules $modules $options (wd: /etc/init.d) [ÃCHEC ] I can reproduce this on i586, but the problem is not present on x86-64 (both Mageia Cauldron). I wonder what causes it. Regards, -- Shlomi Fish
CC: (none) => shlomif
Here is what I discovered using gdb: <QUOTE> [root@telaviv1 ~]# gdb --command=cmds.gdb /usr/sbin/httpd GNU gdb (GDB) 7.1-6.mga1 (Mageia release 1) Copyright (C) 2010 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "i586-mageia-linux-gnu". For bug reporting instructions, please see: <http://www.gnu.org/software/gdb/bugs/>... Reading symbols from /usr/sbin/httpd...Reading symbols from /usr/lib/debug/usr/sbin/httpd.debug...done. done. [Thread debugging using libthread_db enabled] warning: no loadable sections found in added symbol-file /usr/lib/debug/usr/lib/apache-extramodules/mod_perl.so.debug warning: no loadable sections found in added symbol-file /usr/lib/debug/usr/lib/perl5/5.14.1/i386-linux-thread-multi/CORE/libperl.so.debug Program received signal SIGSEGV, Segmentation fault. 0xb78a8a84 in modperl_env_clear (my_perl=0x821d728) at modperl_env.c:100 100 HV *hv = ENVHV; (gdb) print my_perl $1 = (PerlInterpreter *) 0x821d728 (gdb) print PL_curinterp $2 = (PerlInterpreter *) 0x821d728 (gdb) print PL_curinterp->Ienvgv->sv_u.svu_gp->gp_hv[0] $3 = {sv_any = 0x8225f38, sv_refcnt = 1, sv_flags = 549453836, sv_u = { svu_pv = 0x823cd60 "", svu_iv = 136564064, svu_uv = 136564064, svu_rv = 0x823cd60, svu_array = 0x823cd60, svu_hash = 0x823cd60, svu_gp = 0x823cd60, svu_fp = 0x823cd60}} (gdb) print my_perl->Ienvgv->sv_u.svu_gp->gp_hv[0] Cannot access memory at address 0xd (gdb) print PL_curinterp == my_perl $4 = 1 (gdb) print sizeof(my_perl[0]) $5 = 1752 (gdb) print sizeof(PL_curinterp[0]) $6 = 1768 (gdb) </QUOTE> Apparently, there's an inconsistency between the PerlInterpreter definition in those places. This may be caused by inconsistent compiler flags.
Could someone fix this problem? Having the Apache server unable to start seems pretty critical to me. For now, the workaround is to uninstall mod_perl.
Assignee: bugsquad => dmorganec
Looks like this problem happens with Fedora too.
See Also: (none) => https://bugzilla.redhat.com/show_bug.cgi?id=730832
Shlomi, could you have a look at https://bugzilla.redhat.com/show_bug.cgi?id=730832#c8 ?
(In reply to comment #5) > Shlomi, could you have a look at > https://bugzilla.redhat.com/show_bug.cgi?id=730832#c8 ? OK, I will. Right now I'm upgrading a snapshot of my i586 Mageia 1 VM to Mageia Cauldron so I can test it there. It will be time-consuming, so I'll keep you posted.
(In reply to comment #5) > Shlomi, could you have a look at > https://bugzilla.redhat.com/show_bug.cgi?id=730832#c8 ? OK, I tested it and it seems to fix the problem. Please test apache-mod_perl-2.0.5-12.mga2 (which incorporates the patch in question) and let me know if it fixes the problem for you.
Pinging, because nothing has happened with this report for more than 3 months, it still has the status NEW or REOPENED. @ D Morgan Please set status to ASSIGNED if you think this bug was assigned correctly. If for work flow reasons you can't do that, then please put OK on the whiteboard instead. Don't change anything if you want to be pinged by me again in this report;)
CC: (none) => marja11
I'm resolving this bug report as fixed. If Mr. Buclin is still experiencing problems, he is free to reopen this bug report, and I'll deal with the new problems further. But this bug is working for me.
Status: NEW => RESOLVEDResolution: (none) => FIXED