How Zone Files Are Read

[Reindl Harald]

Am 16.12.20 um 18:26 schrieb Gregory Sloop:
> This isn't, IMO, very useful as a response to the OP.

let that decide the OP

> To sum up the response; "It's better to never fail!"
> 
> Yes, that seems pretty obvious. It *would* be better to never fail. Way, 
> way better.
> But the big problem in life is; We're always failing! Dammit!
> 
> So, learning how to gracefully fail, and understanding what happens and 
> why, when something fails, is pretty important to achieve the outcome 
> of; "Not failing quite so catastrophically."

loading a invalid zoen file is far away from "fail geraceful"! if a 
comozter don't understand the input fully it's not supposed to guess

> So, while I don't have helpful knowledge to impart to the OP, I think I 
> can say that giving the advice of "don't fail" doesn't seem very helpful.

where did i give the advice "don't fail"?
please read my repsonse again!

* the zone fails on the master
* the zone is still available on the slaves
* so the error isn't fatal
* but you recognize your mistake

what happens when the error is in the line of the MX record and named 
would say "well, it's only one line, we still have the zone but no 
longer an MX"?

it would lead to a *fatal error* for the behavior of the whole zone, 
even if *all* or your nameservers go down it would be better because 
every delivering MTA would just queue the messages in case of a SERVFAIL

without the MX the would go to the A record of the zone which is in most 
cases simply the wrong destination

> *RH> Am 16.12.20 um 17:37 schrieb Tim Daneliuk:
>>> I ran into a situation yesterday which got me pondering something about bind.
> 
>>> In this case, a single line in a zone file was bad.  The devops automation
>>> had inserted a space in the hostname field of a PTR record.
> 
>>> What was interesting was that - at startup - bind absolutely refused
>>> to load the zone file at all.  I would have expected it to complain
>>> about the bad record and ignore it, but load the rest of the
>>> good records.
> 
>>> Can someone please explain the rationale or logic for this?  Not complaining,
>>> just trying to understand for future reference.
> 
> RH> it's better not load a invalid zone on a single nameserver at all as you
> RH> are supposed to have at least two nameservers and the second one won't
> RH> get the failure via master/slave replication
> 
> RH> if it has an error something is wrong
> RH> if the last version had no error that version is good
> 
> RH> for the world *everything* still is good as long there is one slave -
> RH> subtle errors can lead to completly unexpected behavior