On 17 Feb 2018, at 10:44, Paul paul@roteserver.de wrote:
I am running two relays on debian (one on Raspian 9 and one on Debian 9) and one of them for over three months now.
Are you running an exit?
Please link to your relays on Relay Search.
Some time ago the older one started to loop at startup.
I reinstalled tor and it worked for some time only to loop again.
The last notices entry looks like this "Feb 16 19:36:14.000 [notice] Bootstrapped 0%: Starting" and the next is "Feb 16 19:42:08.000 [notice] Tor 0.2.9.14 (git-a211f886ad759cab) opening log file." then it tells about opening the listeners and parsing the GEOIP files, then my relay identity and the bootstrapped 0% entry again (left some informational entries out).
What is the log entry right before the relay restarts?
Please don't leave out log entries. When you do, we have to guess what the missing log entries are.
If you think the log is too long, put it in a pastebin that can be accessed using Tor Browser in high security mode.
The debug log looks like it is getting descriptors and stuff correctly.
What could be the problem?
Are you using systemd, or some other startup notification system? Your relay could be failing to notify systemd that it has started. And then systemd is restarting it.
Are there limits on RAM, file descriptors, or disk usage on your system? Your relay could be exceeding some kind of OS limit, and being terminated. And then systemd is restarting it.
I'd really need to see your logs to work out what's happening.
T