[Dnsmasq-discuss] [Patch] dnsmasq biased round robin for new requests

Simon Kelley simon at thekelleys.org.uk
Mon Jun 6 12:03:02 BST 2011


harish badrinath wrote:
> Hello,
> 
>> How is your patch better? Maybe because  it avoids the command-line
>> server even if it is fastest, unless the others are down? I can see that
>> would be useful sometimes.
> 
> I would not say better, but you summed it up almost perfectly.
> So back to my original point of contention,is it worthy of being included
> in dnsmasq ??

There are two aspects to that. Is the patch good, and is it a good thing
to have?

The patch needs some changes: the best way to do what you want is to
make changes to reply_query() so that when it gets the first reply from
a low-rank server, it doesn't free the forwarding record, thus allowing
a high-rank server which replies later to become the prefered server. It
might be necessary to add a flag to the forwarding record to stop the
second reply being sent back to the original requestor, or that might be OK.

Is it a good thing to have? That's more difficult to answer. One test is
"can you explain clearly what this configuration option does and what
it's for" ie, write the patch to the man page. Another test is to ask
yourself is "will anyone ever use this option apart from me". That
obviously depends on the first test: no-one will use it if they can't
understand what it does. Assuming they understand it, is it generally
useful? That's judgment call.  Final test is to ask the members of the
mailing list for their opinions, that's one of the things it's there
for, so I throw this open to the floor, what do the dnsmasq users here
think? Mike E, having clarified what the patch does, is it still
potentially useful to you?

Cheers,

Simon.



> 
> Also, thank you for dnsmasq.
> 
Thanks. It's fun to do.

Cheers,

Simon.




More information about the Dnsmasq-discuss mailing list