Red Hat Bugzilla – Bug 1466441
ns-slapd: ../../../include/k5-thread.h:384: k5_mutex_lock: Assertion `r == 0' failed.
Last modified: 2018-04-03 17:35:27 EDT
Description of problem:
This morning ns-slapd crashed with:
Jun 29 08:58:28 europa ns-slapd: ns-slapd: ../../../include/k5-thread.h:384: k5_mutex_lock: Assertion `r == 0' failed.
Version-Release number of selected component (if applicable):
Just once so far
Also, I think /usr/lib/systemd/system/dirsrv@.service would benefit from the addition of:
That crash is in mit krb, not directory server. I'm going to re-assign the problem to them.
Additionally, we have talking about restart on-failure, and we have chosen not to use it due to concerns about systemd and issues like this.
Happened again last evening:
Jun 29 17:26:25 europa ns-slapd: ns-slapd: ../../../include/k5-thread.h:384: k5_mutex_lock: Assertion `r == 0' failed.
Seems like restart on-failure would have at least got my ldap server back up and running instead of being down all evening. I'll add it to /etc/sysconfig/dirsrv.systemd and see if it helps.
Jun 29 04:21 kernel 3.10.0-514.26.1.el7
Jun 20 03:49:59 Updated: glibc.x86_64 2.17-157.el7_3.4
Looks like the trigger was the clock being ~281 seconds off.
k5_mutex_lock is a a very thin wrapper around pthread_mutex_lock that calls assert() if the latter returns nonzero. Please 1) update to rhel-7.4 and then 2) provide me a full backtrace (coredump is preferred, but I understand if you don't want to give me that).
*** Bug 1478619 has been marked as a duplicate of this bug. ***