Red Hat Bugzilla – Bug 808597
sssd_nss crashes on request when no back end is running
Last modified: 2012-06-20 07:56:39 EDT
This bug is created as a clone of upstream ticket: https://fedorahosted.org/sssd/ticket/1270 uname -r 3.2.10-3.fc16.x86_64 After a few minutes when sssd is launched : systemctl restart sssd.service Process are : {{{ 5628 ? Ss 0:00 /usr/sbin/sssd -D -f 5630 ? S 0:00 /usr/libexec/sssd/sssd_nss --debug-to-files 5631 ? S 0:00 /usr/libexec/sssd/sssd_pam --debug-to-files 5663 ? R 0:00 /usr/libexec/sssd/sssd_be --domain LDAP --debug-to-files }}} The log says : {{{ [root@lame4: 382] Mar 20 18:07:36 lame4 kernel: [11304.391716] sssd_be[5663]: segfault at 4000 ip 00007fc4188e64a2 sp 00007ffffbf0ce90 error 4 in sssd_be[7fc41888d000+75000] Mar 20 18:07:36 lame4 abrtd: Directory 'ccpp-2012-03-20-18:07:36-5663' creation detected Mar 20 18:07:36 lame4 abrt[5682]: Saved core dump of pid 5663 (/usr/libexec/sssd/sssd_be) to /var/spool/abrt/ccpp-2012-03-20-18:07:36-5663 (16244736 bytes) Mar 20 18:07:36 lame4 sssd[be[LDAP]]: Starting up Mar 20 18:07:36 lame4 abrtd: DUP_OF_DIR: /var/spool/abrt/ccpp-2012-03-20-15:04:40-828 Mar 20 18:07:36 lame4 abrtd: Problem directory is a duplicate of /var/spool/abrt/ccpp-2012-03-20-15:04:40-828 Mar 20 18:07:36 lame4 abrtd: Deleting problem directory ccpp-2012-03-20-18:07:36-5663 (dup of ccpp-2012-03-20-15:04:40-828) }}} Process are : {{{ 5628 ? Ss 0:00 /usr/sbin/sssd -D -f 5630 ? S 0:00 /usr/libexec/sssd/sssd_nss --debug-to-files 5631 ? S 0:00 /usr/libexec/sssd/sssd_pam --debug-to-files }}} A few minutes later log says : {{{ Mar 20 18:29:01 lame4 abrtd: Directory 'ccpp-2012-03-20-18:29:01-5630' creation detected Mar 20 18:29:01 lame4 abrt[6320]: Saved core dump of pid 5630 (/usr/libexec/sssd/sssd_nss) to /var/spool/abrt/ccpp-2012-03-20-18:29:01-5630 (1622016 bytes) Mar 20 18:29:01 lame4 sssd[nss]: Starting up Mar 20 18:29:01 lame4 sssd[nss]: Starting up Mar 20 18:29:01 lame4 sssd[nss]: Starting up Mar 20 18:29:01 lame4 abrtd: DUP_OF_DIR: /var/spool/abrt/ccpp-2012-03-20-15:11:34-833 Mar 20 18:29:01 lame4 abrtd: Problem directory is a duplicate of /var/spool/abrt/ccpp-2012-03-20-15:11:34-833 Mar 20 18:29:01 lame4 abrtd: Deleting problem directory ccpp-2012-03-20-18:29:01-5630 (dup of ccpp-2012-03-20-15:11:34-833) }}} Process are : {{{ 5628 ? Ss 0:00 /usr/sbin/sssd -D -f 5631 ? S 0:00 /usr/libexec/sssd/sssd_pam --debug-to-files }}} and getent passwd user doesn't work anymore In a first time sssd_be segfaults and getent passwd user works, ans in a second time sssd_nss try to start up and getent passwd user fails.
This bugzilla is tracking the sssd_nss problem. The sssd_be crash is being tracked separately in bug #808597. To reproduce: 1. kill sssd_be so that is stops definitely for x in $(seq 1 4); do kill -9 $(pidof sssd_be); done 2. make sure no sssd_be process is running now pidof sssd_be 3. getent passwd someuser With the current packages, the sssd_nss process crashes with SIGABRT. It should not crash, but rather revert to returning data from cache.
# getent passwd puser1 puser1:*:1001:1001:Posix User1:/home/puser1: # for x in $(seq 1 4); do kill -9 $(pidof sssd_be); done # pidof sssd_be # getent passwd puser1 puser1:*:1001:1001:Posix User1:/home/puser1: Verified with sssd-1.8.0-23
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: No documentation required
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHBA-2012-0747.html