[tor-talk] Tor service can't start with "CPU worker" error

[Roger Dingledine]

On Sat, Aug 27, 2011 at 05:05:06PM -0400, Advrk Aplmrkt wrote:
> Hello,
> I've been running a Tor relay for almost 5 days now and it works great!
> Unfortunately, I rebooted my server (which is running Tor) today, and
> when I do "service tor status", I get the following:
> 
> PID file (/var/run/tor/tor.pid) present, but tor (1914) not running
> 
> I tried:
> 
> (1) Rebooting the server again
> 
> (2) running "service tor restart"
> 
> But that didn't help. So I looked at the logs (info level), and found
> the following on the last lines:
> 
> Aug 27 16:53:24.447 [info] choose_good_exit_server_general(): Chose
> exit server 'TOR1CINIPAC'
> Aug 27 16:53:24.452 [info] circuit_handle_first_hop(): Next router is
> [scrubbed]: Connection in progress; waiting.
> Aug 27 16:53:24.452 [info] cpuworker_main(): CPU worker exiting
> because Tor process closed connection (either rotated keys or died).
> 
> There seems to be a problem with the "cpuworker"? But what does that mean?

It is an "info" level log. It doesn't mean there's a problem with your
cpuworker. That is a normal log line, either as part of Tor shutting down,
or as part of Tor rotating some of its keys. If it's the last line in
your log, my guess is it's part of Tor shutting down.

> By the way, the default notice level tor.log did not show any errors,
> but always end with 80% bootstrap as its last line...
> 
> The Tor I have is version 0.2.1.30 installed via yum from the
> "torproject" RPM repository as indicated on the Tor website. Also, my
> system is Scientific Linux 6.1 64-bit, which is RHEL without the
> brand.
> 
> Can anyone help with this? I am eager to get my relay up and running
> again. Thanks.

Are there no "notice" or "warn" level logs?

If no, see also
https://trac.torproject.org/projects/tor/wiki/doc/TorFAQ#MyTorkeepscrashing.

--Roger