On 06/10/14 10:34, Christian Ruppert wrote: > > > On 10/03/2014 11:02 PM, Simon Kelley wrote: >> On 30/09/14 10:52, Christian Ruppert wrote: >>> Hi Simon, >>> >>> the related code is indeed really simple. I have no idea how that could >>> happen >>> at all. However, I am no longer able to reproduce it :( >>> It might be some corner case or so, idk. I also added some debug logging and >>> watched it in gdb but no luck. I hope it's really gone now, whatever caused >>> it... >>> I'll keep an eye on it. Thanks! >> >> Race between starting NTP and starting dnsmasq? >> > > hrm, time() shouldn't return 0 or 1 then anyway. > I am currently no longer able to reproduce the issue. I restarted dnsmasq > several times etc. all fine for now. Even starting without NTP or so doesn't > matter.
It's a mystery. > > I also noticed that dnsmasq increases the serial even if nothing happened at > all > (at least nothing DNS related). But this is probably another issue. It bumps the serial whenever a change happens to the DHCP lease database that _might_ change the set of DNS records that come from DHCP. It's not totally stupid (like, it shouldn't increment when an existing lease is renewed for the same address and hostname. But there will certainly be cases when the serial is bumped but no actual change occurs in the DNS. Cheers, Simon. > >> _______________________________________________ Dnsmasq-discuss mailing list Dnsmasq-discuss@lists.thekelleys.org.uk http://lists.thekelleys.org.uk/mailman/listinfo/dnsmasq-discuss
