KEA-DHCP6 fail to start at some point
-
At some point, not sure exactly when it starts to happening, the KEA-DHCP6 fails to start.
/usr/local/etc/kea/kea-dhcp6.conf shows as empty fileERROR [kea-dhcp6.dhcp6.0x2920d8412000] DHCP6_INIT_FAIL failed to initialize Kea server: configuration error using file '/usr/local/etc/kea/kea-dhcp6.conf': Error initializing hooks: CmdHttpListener::run failed: unable to setup TCP acceptor for listening to the incoming HTTP requests: bind: Address already in use [system:48 at /usr/local/include/boost/asio/detail/reactive_socket_service.hpp:161:33 in function 'bind']
Any thoughts?
-
Does it start after that if you try manually? Or resave the dhcpv6 settings?
-
@stephenw10
No, only pfSense restart did the trick. -
@w0w said in KEA-DHCP6 fail to start at some point:
ERROR [kea-dhcp6.dhcp6.0x2920d8412000] DHCP6_INIT_FAIL failed to initialize Kea server: configuration error using file '/usr/local/etc/kea/kea-dhcp6.conf': Error initializing hooks: CmdHttpListener::run failed: unable to setup TCP acceptor for listening to the incoming HTTP requests: bind: Address already in use [system:48 at /usr/local/include/boost/asio/detail/reactive_socket_service.hpp:161:33 in function 'bind']
Any thoughts?
"http", so some kea or sub kea process that uses a TCP port ?
sockstat | grep 'kea'
... nope.
Even stranger : You saw where the file was ? /usr/local/include/boost/asio/detail/ !
That's in /usr/local/include/ .... I always thought that place was reserved for header files (pretty useless for pfSense as it is not a dev system) ... Well, hpp is a c++ header file. But these executable or usable ?
I could find a 'kea' process using some port serving http requests ...What happened - imho - was : some kea dhcp6 sub process became a zombie, so the OS could stop it anymore. A port got bound permanently, and when dhcp6 rstarted, (a ?) process could reuse it as it was already bound: hence the error.
Next time, if you see this very rare error again : before restarting the entire pfSense, look what process this is ?
This zombie process could probably be removed with the "kill" command on the command line. -
@Gertjan
Yes, it looks like the issue is indeed with the zombie process. If it happens again, I'll definitely check. Thanks for the idea.