|Summary:||samba 4.6.2-11.el7_4: unreliable winbindd start with Type=notify and double forking|
|Product:||Red Hat Enterprise Linux 7||Reporter:||Alexey Dobriyan <adobriyan>|
|Component:||samba||Assignee:||Andreas Schneider <asn>|
|Status:||CLOSED ERRATA||QA Contact:||Andrej Dzilský <adzilsky>|
|Version:||7.6-Alt||CC:||adzilsky, asn, gdeschner, ian.donaldson, jrivera, lmiksik, peter.clark, rhack, toneata, tscherf|
|Fixed In Version:||samba-4.7.1-4.el7||Doc Type:||If docs needed, set a value|
|Doc Text:||Story Points:||---|
|:||1686048 (view as bug list)||Environment:|
|Last Closed:||2018-04-10 17:30:15 UTC||Type:||Bug|
|oVirt Team:||---||RHEL 7.3 requirements from Atomic Host:|
|Cloudforms Team:||---||Target Upstream Version:|
|Bug Depends On:|
Description Alexey Dobriyan 2017-11-08 11:15:31 UTC
Description of problem: winbindd is shipped with the following .service file: Type=notify PIDFile=/run/winbindd.pid ExecStart=/usr/sbin/winbindd "$WINBINDOPTIONS" Default WINBINDOPTIONS are empty string. Internally winbindd does double fork and notifies systemd about main process PID change at lib/util/become_daemon.c:become_daemon() Apparently this confuses systemd as Type=notify daemons aren't supposed to double fork. We got report from user which reboots system often and every once in a while winbindd doesn't actually start despite reporting OK status. Manual restart fixes the problem. We got confirmation that using WINBINDOPTIONS="-F --no-process-group" (and deleting PIDFile line for a good measure) makes winbindd startup reliable. Version-Release number of selected component (if applicable): samba-4.6.2-11.el7_4.src.rpm et al How reproducible: rarely Steps to Reproduce: 1. reboot machine until winbindd doesn't start Actual results: Expected results: winbindd starts 100% reliably Additional info:
Comment 2 Alexey Dobriyan 2017-11-08 11:21:00 UTC
smbd and nmbd should have the same problem as they call into become_daemon() as well.
Comment 3 Alexey Dobriyan 2017-11-08 11:22:21 UTC
> Additional info: What happens is that systemd sends spurious SIGTERM in the middle of daemon start.
Comment 6 Andrej Dzilský 2017-12-18 13:55:21 UTC
This rare bug is Hard to even reproduce nor test. Sanity Only.
Comment 10 errata-xmlrpc 2018-04-10 17:30:15 UTC
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. https://access.redhat.com/errata/RHBA-2018:0937