LS,
Our FQDND process crashed. We are using Consul DNS to create hostnamed NAT policies to our VMs. It works very well, which is great. However, we stumbled upon a bug in FQDND.
MESSAGE Apr 28 22:46:23 [4157512512]: execute_fqdn_opcode:call do_opcode for FQDN Host : traefik.service.consul
MESSAGE Apr 28 22:46:25 [4158342912]: execute_fqdn_opcode:call do_opcode for FQDN Host : traefik.service.consul
MESSAGE Apr 28 22:46:35 [4157512512]: execute_fqdn_opcode:call do_opcode for FQDN Host : traefik.service.consul
MESSAGE Apr 28 22:46:36 [4158342912]: execute_fqdn_opcode:call do_opcode for FQDN Host : traefik.service.consul
MESSAGE Apr 28 22:57:06 [4157512512]: execute_fqdn_opcode:call do_opcode for FQDN Host : traefik.service.consul
MESSAGE Apr 28 22:57:06 [4157512512]: execute_fqdn_opcode:call do_opcode for FQDN Host : firewall-traefik.service.consul
MESSAGE Apr 28 22:57:07 [4158342912]: execute_fqdn_opcode:call do_opcode for FQDN Host : firewall-traefik.service.consul
free(): invalid next size (normal)
Our FQDND process crashed, and we had to restart it via the advanced shell. When I got there, I captured the above lines from the logs (fqdndebug.log being empty). For your info, the hostname firewall-traefik has a TTL of 2s, and traefik.service.consul has 0s TTL. I guess I initially created the wrong hostname (without the firewall prefix) in the interface, and then switched to the correct hostname (with prefix). In the firewall we want to use a 2 sec TTL because we noticed with 0s TTL many DNS requests per second are executed. I have no idea why FQDND crashes at "invalid next size".
Hopefully you can figure out the problem and solve in a newer version.
Regards,
Frederik
This thread was automatically locked due to age.
