The newly introduced Mit-krb5 does not work with samba v4.7.0 on Centos7.X-x86-64

fu-hong-quan at fu-hong-quan at
Wed Sep 27 06:59:17 UTC 2017


  I'm running samba v4.7.0 as AD DC with Mit-krb5, yet krb5kdc exits soon 
after a joined windows7-64bit starts up, with the screen of prompting 

 to press "CTRL+ATL+DELETE" to type user account.
 The symptom is that krb5 exit with status 6, and coredump happens, see 

 [root at pdc samba]# more log.samba
[2017/09/27 12:54:55.362600,  0] 
  /usr/sbin/krb5kdc: krb5kdc: do_tgs_req.c:826: process_tgs_req: Assertion 
`status != ((void *)0)' failed.
[2017/09/27 12:54:55.833120,  0] 
  The MIT KDC daemon died with exit status 6
[2017/09/27 12:54:55.833268,  0] 
  task_server_terminate: [mitkdc child process exited]

Part of coredump info is:

Wed Sep 27 12:54 1117/87310 "[abrt] krb5-server: krb5kdc killed by 

cmdline:        /usr/sbin/krb5kdc -n
executable:     /usr/sbin/krb5kdc
package:        krb5-server-1.15.1-8.el7
component:      krb5

:Sep 27 11:55:56 samba[14387]:   /usr/sbin/krb5kdc: 
krb5kdc: do_tgs_req.c:826: process_tgs_req: Assertion `status != ((void 
*)0)' failed.
:Sep 27 11:55:56 abrt-hook-ccpp[14593]: Process 14394 
(krb5kdc) of user 0 killed by SIGABRT - dumping core

I tries on centos7.X-x86-64(X=7.2 and 7.4) for several times, it happens 
every time.

And kadmin.local command also coredumpped as long as I run "listprincs" 
and "exit"

While the embedded Hemedal works quite well. 

So is it a bug?

More information about the samba-technical mailing list