Description of problem: Openldap 2.4.23 acting as authentication source for small server with cca 500 accounts. slapd crashes after several hours or days. "/var/log/messages" contain items as these: Jun 1 13:39:32 joko kernel: [97730.671115] slapd[26759]: segfault at 66fd40b4 ip 0012bbc8 sp b4dfddf0 error 6 in libssl3.so[110000+34000] Jun 14 14:51:34 joko kernel: [133763.258801] slapd[10157]: segfault at 679d40b4 ip 00163d18 sp b4cfddf0 error 6 in libssl3.so[148000+34000] Jun 20 09:53:30 joko kernel: [272223.479308] slapd[26602]: segfault at 67dd40b4 ip 00739d18 sp b51fcdf0 error 6 in libssl3.so[71e000+34000] Jun 22 12:21:11 joko kernel: [453884.312551] slapd[23588]: segfault at 667d40b4 ip 00950d18 sp b53fedf0 error 6 in libssl3.so[935000+34000] Jun 24 11:24:28 joko kernel: [19209.436001] slapd[1685]: segfault at 20 ip 00b84f4d sp b52fe7c4 error 4 in libpthread-2.13.so[b7d000+17000] Jun 27 10:16:48 joko kernel: [274349.615288] slapd[2858]: segfault at 669d40b4 ip 00165d18 sp b54fedf0 error 6 in libssl3.so[14a000+34000] Jul 11 12:12:21 joko kernel: [622636.907118] slapd[30524]: segfault at 675d40b4 ip 0047dd18 sp b58fedf0 error 6 in libssl3.so[462000+34000] Version-Release number of selected component (if applicable): openldap-2.4.23-10.fc14.i686 How reproducible: probably not possible, seems as it crashes randomly. Additional info: Problem appears after upgrade to F14. Previous openldap-2.4.12@Fedora 10 at same machine was worked without any problems.
Not sure where openldap dumps core. You'll have to enable core dumps for setuid programs first: sysctl -w fs.suid_dumpable=1 then edit /etc/sysconfig/ldap and add this: ulimit -c unlimited then install the openldap-debuginfo package debuginfo-install openldap make sure this installs the nss-debuginfo package as well Then service ldap restart Then attempt to reproduce the problem Not sure if this will work - if it does, not sure where the core file will be generated Once you have a core dump, do this gdb /usr/sbin/slapd /path/to/core.PID (gdb) set logging on (gdb) set logging file /tmp/slapd.bt.txt (gdb) set logging redirect on (gdb) thread apply all bt full (gdb) set logging redirect off (gdb) quit EDIT /tmp/slapd.bt.txt TO REMOVE/OBSCURE ANY SENSITIVE DATA then attach /tmp/slapd.bt.txt to this bug as an attachment (text/plain)
(In reply to comment #1) > Not sure where openldap dumps core. The easiest way is to get ABRT installed and enabled. The core dumps are then collected in /var/spool/abrt
I have the same problem with openldap-2.4.23-15.el6 on rhel6.1
(In reply to comment #3) > I have the same problem with openldap-2.4.23-15.el6 on rhel6.1 Please install ABRT and generate a core dump, then a stack trace.
When slapd crashes I send the core dump because it crashes randomly.
ping - any more crashes?
I was able to get a core dump on a rhel6 box and submitted it to an existing RHEL6 bug, here: https://bugzilla.redhat.com/show_bug.cgi?id=709407#c8
*** This bug has been marked as a duplicate of bug 701678 ***