[BIND] Why log a failed transfer successfully?

[Hugo Salgado]

On 04/02/2015 05:01 AM, Anand Buddhdev wrote:
> I'm parsing BIND logs to extract the XFR size in bytes of a zone, and
> was just bitten by this sequence:
> 
> 02-Apr-2015 04:27:10.393 xfer-in: transfer of './IN' from
> 2001:67c:2e8:5::c100:c6#53: failed to connect: timed out
> 02-Apr-2015 04:27:10.393 xfer-in: transfer of './IN' from
> 2001:67c:2e8:5::c100:c6#53: Transfer completed: 0 messages, 0 records, 0
> bytes, 62.999 secs (0 bytes/s
> ec)
> 
> So BIND wasn't even able to connect to the master, but subsequently
> logged "Transfer completed".
> 
> Is there any logic to this that I'm missing?

I want to support (if that was Anand's original idea) a better log line
in this case. We host several thousand zones as secondary service, and
provide log access to customers. I estimate a couple of questions per
week in our helpdesk about people assuming that the transfers are
working, overlooking exactly that line.

Thanks,

Hugo