creating IPv6 interface eth0 failed; interface ignored

[Sotiris Tsimbonis]

On 19/8/16 12:38, Mukund Sivaraman wrote:
> On Fri, Aug 19, 2016 at 11:32:43AM +0200, Wolfgang Riedel wrote:
>> ### bootup with: empty-zones-enable no;
>>
>> [root at ns1 ~]# systemctl status named-chroot.service
>> ● named-chroot.service - Berkeley Internet Name Domain (DNS)
>>    Loaded: loaded (/usr/lib/systemd/system/named-chroot.service; enabled; vendor preset: disabled)
>>    Active: active (running) since Sat 2016-08-06 11:08:22 CEST; 16s ago
>>   Process: 1084 ExecStart=/usr/sbin/named -u named -t /var/named/chroot $OPTIONS (code=exited, status=0/SUCCESS)
>>   Process: 1080 ExecStartPre=/bin/bash -c if [ ! "$DISABLE_ZONE_CHECKING" == "yes" ]; then /usr/sbin/named-checkconf -t /var/named/chroot -z /etc/named.conf; else echo "Checking of zone files is disabled"; fi (c
>>  Main PID: 1086 (named)
>>     Tasks: 5 (limit: 512)
>>    CGroup: /system.slice/named-chroot.service
>>            └─1086 /usr/sbin/named -u named -t /var/named/chroot
>>
>> Aug 06 11:08:22 ns1.f1-online.net named[1086]: listening on IPv6 interface lo, ::1#53
>> Aug 06 11:08:22 ns1.f1-online.net named[1086]: listening on IPv6 interface eth0, 2001:67c:21b0:4029:193:34:29:244#53
>> Aug 06 11:08:22 ns1.f1-online.net named[1086]: could not listen on UDP socket: address not available
>> Aug 06 11:08:22 ns1.f1-online.net named[1086]: creating IPv6 interface eth0 failed; interface ignored
> 
> Assuming this the broken state you're describing (as you've attached
> before and after log copies), from the log messages above it seems the
> interface is not available when named is being started.
> 
> I have seen this behavior with several other services on Fedora that
> need manual restart after boot (e.g., postfix, nginx and sshd) to make
> them listen on all configured interfaces because the interface was not
> configured when the service was being started.
> 

A workaround for this behavior on linux systems is

  echo "net/ipv4/ip_nonlocal_bind=1" >> /etc/sysctl.conf

Then daemons won't fail when systemd starts them and will listen to the
configured ip address when it becomes available (tested with nginx on
rhel7).

Kind regards,
Sotiris.