Bug 13230 - BIND9 DLZ backend needs workaround
BIND9 DLZ backend needs workaround
Status: NEW
Product: Samba 4.1 and newer
Classification: Unclassified
Component: Documentation
4.7.4
All Linux
: P5 normal
: ---
Assigned To: Samba Documentation QA Contact~
Samba Documentation QA Contact~
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2018-01-18 11:19 UTC by Kristján Jónsson
Modified: 2018-02-01 10:42 UTC (History)
0 users

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Kristján Jónsson 2018-01-18 11:19:27 UTC
A problem exists with the samba_dlz in bind9 on certain distros.
It has been observed on CentOS 7, which uses systemd.

a 'reload' of named causes dlz_samba.so to be destroyed.

This can happen as part of log rotation.  the file /etc/logrotate.d/named contatins a "systemctl reload named.service".

If this is the case, the user should modify this file to do:
"systemctl restart named.service".

The user can check for this service by performing
"systemctl reload named.service", then doing a "journalctl -e" and looking for a line similar to:

"Jan 12 09:58:17 host.example.net named[8331]: samba_dlz: shutting down" indicates that this may be a problem on this server.




I think that a clause should be added to https://wiki.samba.org/index.php/BIND9_DLZ_DNS_Back_End to explain this.

Assuming that this is not just a problem on systemd, the proper check is to:
1) perform 'rndc reload'
2) examine the syslog or journal, look for "samba_dlz: shutting down"
3) if it happens, examin /etc/logrotate/named and see if a 'reload' needs to be turned into 'restart'
4) restart bind with 'rndc restart'
Comment 1 Kristján Jónsson 2018-02-01 10:42:15 UTC
Okay, https://wiki.samba.org/index.php?title=BIND9_DLZ_DNS_Back_End&oldid=14154
did add a mention that "reload" doesn't work.  It failed to mention however the most common source of this reload, /etc/logrotate.d/named.