Red Hat Bugzilla – Bug 961908
[abrt] realmd-0.14.0-1.fc19: exit: Process /usr/lib64/realmd/realmd was killed by signal 6 (SIGABRT)
Last modified: 2013-05-13 16:06:02 EDT
Description of problem:
I was in the middle of performing test 'https://fedoraproject.org/wiki/QA:Testcase_realmd_join_sssd?rd=QA:Testcase_Active_Directory_realmd_join_sssd'
I had just issued 'sudo klist -k'
Version-Release number of selected component:
runlevel: N 5
Thread no. 1 (1 frames)
#9 exit at exit.c:99
Created attachment 746284 [details]
Created attachment 746285 [details]
Created attachment 746286 [details]
Created attachment 746287 [details]
Created attachment 746288 [details]
Created attachment 746289 [details]
Created attachment 746290 [details]
Created attachment 746291 [details]
Created attachment 746292 [details]
Created attachment 746293 [details]
This happened when realmd was exiting (due to its timeout). The crash happened in libselinux. realmd makes no explicit selinux calls, and this seems to be some sort of atexit() or dl_fini handler.
libselinux has had this behaviour for ever. Usually when an error happens it is caused by a memory problem. Basically libselinux at load allocates some memory and at unload frees the memory. If something in the app has written over memory then it can cause the crash. Can we run realmd through valgrind to see if it has some memory problems.
Makes sense. Then this is almost certainly a duplicate of bug 953486
*** This bug has been marked as a duplicate of bug 953486 ***