Description of problem:

During SMB authentication the main cronie process terminates so no other jobs are executed after a single auth fault.

Before 1.4.7 and patch

https://github.com/cronie-crond/cronie/commit/ee4cbe7659ede3f61db18cc922ee0e27268a8579?diff=unified

cronie was known to be overly sensitive to SIGPIPE.

Version-Release number of selected component (if applicable):

cronie-1.4.4-16.el6_8.2

How reproducible:

Often in customer's environment

Steps to Reproduce:
1. setup crontab for user in SMB domain
2. observe SMB auth failure

Actual results:

main cronie process terminates on SIGPIPE, no record in logs

Expected results:

cronie logs the auth failure and recovers, continues serving subsequent jobs

Additional info:

messages events preceding the cronie crash:

[2018/07/29 19:39:09.683477,  0] auth/pampass.c:589(smb_pam_account)
  smb_pam_account: PAM: UNKNOWN PAM ERROR (4) during Account Management for User: netadmin
[2018/07/29 19:39:09.683612,  0] auth/pampass.c:797(smb_pam_accountcheck)
  smb_pam_accountcheck: PAM: Account Validation Failed - Rejecting User netadmin!
[2018/07/29 19:39:09.684137,  0] rpc_server/srv_pipe.c:835(pipe_gssapi_verify_final)
 rpc_server/srv_pipe.c:835: failed to obtain the server info for authenticated user: NT_STATUS_ACCESS_DENIED
[2018/07/29 19:39:09.684270,  0] rpc_server/srv_pipe.c:1608(api_pipe_alter_context)

Jul 29 19:39:01 server CROND[26999]: (smbuser) CMD (some cmd)
Jul 29 19:52:28 server crond[3308]: (CRON) STARTUP (1.4.4)             (manual)

in cronie strace:

...
connect(7, {sa_family=AF_LOCAL, sun_path="/var/lib/samba/winbindd_privileged/pipe"}, 110) = 0 <0.000132>
...
write(7, "0\10 ...0\0\0smbuser\0\0\...
--- SIGPIPE