Recently I upgraded a server from Fedora 21 to Fedora 23. This was a
difficult process because newer kernels (> 3.14.23-100.fc19) won’t work
with the system.
Finally I was able to boot the system with the old kernel, but I was not
seeing any information in /var/log/maillog.
Sendmail was running and I was able to send and receive messages, but
nothing was written to the maillog.
When I talked to somebody in #fedora he mentioned I could do a test by
running the following simple command:
When the Fedora 20 server at my parents rebooted because of a general
power failure, the tinc tunnel between my parents and myself didn’t work
anymore.
Troubleshooting
I enabled the debug logging by sending an INT signal to the daemon.
tincd -n ubbink -k INT
This resulted in the following log items (journalctl -f -l –unit
tincd@ubbink.service):
tinc.ubbink[9707]: Got INT signal
tinc.ubbink[9707]: Temporarily setting debug level to 5. Kill me with SIGINT again to go back to level 0.
tinc.ubbink[9707]: Trying to connect to amys (xxx.xxx.xxx.xxx port 655)
tinc.ubbink[9707]: xxx.xxx.xxx.xxx port 655: Network is unreachable
tinc.ubbink[9707]: Could not set up a meta connection to amys
tinc.ubbink[9707]: Trying to re-establish outgoing connection in 35 seconds
tinc.ubbink[9707]: Purging unreachable nodes
I couldn’t find why this was happening because when I used netcat to
connect to port 655 via udp everything worked.
nc -vu xxx.xxx.xxx.xxx 655
Also a telnet to port 655 worked, but why wasn’t tinc able to make the
connection?