Slave fails failover

[Jan Engelhardt]

Hello,


as commonly done, I have two DNS servers for a domain, that is, the 
toplevel zone at the hoster is

  $ORIGIN xyz.com.
  jen IN NS dns1.foo.com.  (running ISC BIND 9.3.2)
  jen IN NS dns2.bar.com.  (also ISC BIND 9.3.2)

  $ORIGIN foo.com.
  dns1 IN A 1.2.3.4

dns1's zone file is (shortened):

  $ORIGIN xyz.com.
  @ SOA ...
    IN A 1.2.3.4

dns2 is configured as a slave to dns1. Trying to resolve xyz.com from 
anywhere in the world usually yields 1.2.3.4, which is ok.

Now that dns1 is powered off, I noticed that xyz.com cannot be resolved 
anymore, even though that should have been the purpose of having slave 
servers, is not it?

The dns2 logfile says

Jul 27 11:44:16 dns2 named[5652]: zone jen.xyz.com/IN: refresh: retry
limit for master 1.2.3.4#53 exceeded (source 0.0.0.0#0)
Jul 27 11:44:16 dns2 named[5652]: zone jen.xyz.com/IN: Transfer started.
Jul 27 11:44:19 dns2 named[5652]: transfer of 'jen.xyz.com/IN' from
1.2.3.4#53: failed to connect: host unreachable
Jul 27 11:44:19 dns2 named[5652]: transfer of 'jen.xyz.com/IN' from
1.2.3.4#53: end of transfer

Should not dns2 have kept the zones of the last successful transfer and use 
them?


Jan Engelhardt
--