Just FYI I hit this problem too in the past when trying to set up HA (with some 0.9 beta version). I solved it by setting up a mysql master on every kea node and pointing kea to 127.0.0.1, not optimal but good enough. With high traffic it wouldn't work, though, because of deadlocks and collisions.
Kea should probably support some sort of connection pool with reconnect and keepalive (SELECT 1;), no idea how that works now... Jan > On 2 Jan 2017, at 10:27, Samuel Hameau <[email protected]> wrote: > > Hi, > > I am using Kea with a mysql cluster of 2 master/master nodes using mysql > replication (not a nbd cluster) for leases and hosts reservation. > > For mysql failover, I am using a virtual IP balanced between the 2 mysql > nodes. > Everytime the virtual IP moves to the other host, i got a fatal error, and > kea stops running with that kind of error : > > DHCPSRV_MYSQL_FATAL_ERROR Unrecoverable MySQL error occurred: unable to > execute for <DELETE FROM lease4 WHERE state = ? AND expire < ?>, reason: Lost > connection to MySQL server during query (error code: 2013). Server exiting > now! > > Would it be possible to modify this behavior to make kea retry to connect to > the mysql server a few times before dying ? > > Previously, i have also tried tu use haproxy with one mysql server being > active, and the 2nd one being in backup : I had the same problem when > restarting haproxy/ or changing the active mysql server in haproxy. > > Best regards, > Samuel > _______________________________________________ > Kea-users mailing list > [email protected] > https://lists.isc.org/mailman/listinfo/kea-users
_______________________________________________ Kea-users mailing list [email protected] https://lists.isc.org/mailman/listinfo/kea-users
