Discussion:
[Dnsmasq-discuss] dnsmasq failover
Donald Muller
2018-03-02 17:36:35 UTC
Permalink
At the risk of offending Geert I have a question on failover. I found a thread from 6 years ago discussing dnsmasq failover. There were a number of suggestions made that required enhancements to dnsmasq none of which seem to have been implemented. Is this a dead idea or something that is still on the back (very back) burner?

Thanks
Don
Kurt H Maier
2018-03-02 18:34:20 UTC
Permalink
Post by Donald Muller
At the risk of offending Geert I have a question on failover. I found a thread from 6 years ago discussing dnsmasq failover. There were a number of suggestions made that required enhancements to dnsmasq none of which seem to have been implemented. Is this a dead idea or something that is still on the back (very back) burner?
The simplest approach is to share your configs, usually with a shared
filesystem or drbd if you must, then configure CARP or VRRP, and set up
heartbeat to start up the secondary when the primary fails. This is far
more reliable than trying to juggle which tftp address to pass your pxe
clients, and is generalizable to other services.

khm
Donald Muller
2018-03-03 06:07:17 UTC
Permalink
Thanks Kurt. I don't use tftp so that is not a concern. I was hoping for something a little easier like the two dnsmasq instances talking to each other and passing information. Oh well, c'est la vie. So this is not on the roadmap?

Thanks
Don
-----Original Message-----
Sent: Friday, March 2, 2018 1:34 PM
Subject: Re: [Dnsmasq-discuss] dnsmasq failover
Post by Donald Muller
At the risk of offending Geert I have a question on failover. I found a
thread from 6 years ago discussing dnsmasq failover. There were a number
of suggestions made that required enhancements to dnsmasq none of which
seem to have been implemented. Is this a dead idea or something that is still
on the back (very back) burner?
The simplest approach is to share your configs, usually with a shared
filesystem or drbd if you must, then configure CARP or VRRP, and set up
heartbeat to start up the secondary when the primary fails. This is far more
reliable than trying to juggle which tftp address to pass your pxe clients, and
is generalizable to other services.
khm
Loading...