Hi Anand,
this is a good idea and reasoning, so I'd like to have it implemented.
Maybe not exactly this, but there
should definitely be some mechanism to unstuck slave servers after
operational error.
We're quite busy with preparing the next release, so I can't promise any ETA,
but I have at least started an issue to track this -
https://gitlab.labs.nic.cz/labs/knot/issues/212
Best,
Marek
On 13 February 2014 01:34, Anand Buddhdev <anandb(a)ripe.net> wrote:
Hi Knot developers,
I have a feature request: I'd like a knotc command that will force Knot
to transfer a zone that is configured as a slave. The only command I can
issue at the moment is "refresh", but this will not transfer a zone if
Knot has a higher serial than the master.
Such a forced transfer command is useful to help recover from situations
where due to some operational error on the master server, the serial
number has gone back, and we deliberately need our Knot server to sync
to that old serial number. Without this command, the only way to make
Knot recover is to stop it, delete the zone file from disk and start it
again. But this is not ideal on a running server with lots of zones,
because the stop and start times can be quite high (in our case, 37s to
stop, and 1m30s to start).
Both BIND and NSD provide such commands to ignore the serial number
check and force zone transfers. I'd love to see such a command for Knot too.
Regards,
Anand
_______________________________________________
knot-dns-users mailing list
knot-dns-users(a)lists.nic.cz
https://lists.nic.cz/cgi-bin/mailman/listinfo/knot-dns-users