On 2/25/14, Roger Dingledine arma@mit.edu wrote:
On Tue, Feb 25, 2014 at 10:15:11AM +1100, Zenaan Harkness wrote:
I'm on Debian and did a service tor reload (not restart) and tor crashed! I didn't realise immediately, took may be a minute to realise and restart. Anyway apologies to any connections that were going through this relay.
Should I report a bug to Debian?
First check your logs -- it's likely that it's not a crash, but rather Tor read the new torrc file and either couldn't parse it or refused to switch to the new parameters.
I ran service tor status, and it was definitely dead. What twigged me was rerunning arm and arm began with running some wizard for "new site" configuration, so I knew something had gone wrong, checked the service and sure enough it was "not running".
service tor start and all was back and running within a minute or less, and arm showed me as much.
So definitely tor stopped running - in a way that was deterministic (I did run the service tor reload command) but totally unexpected - tor should never stop running (or crash) with just a config file reload!
Also, this morning, my relay is back up, but no longer with HSDir flag. Is this due to setting that option in the torrc file? And if so, I guess it is recommended to disable that option?
To disable which option? You should leave HidServDirectoryV2 alone in general, and let the Tor directory authorities decide whether you're stable enough to be used.
DisableDebuggerAttachment 0
This is the only change, the one which subsequent to adding that option, and running service tor reload, tor apparently stopped running, as above.
So the only change that I am aware of that could have caused the HSDir flag (as shown in arm) to go away, was either that option (DisableDebuggerAttachment 0), or simply the fact that the relay crashed (or somehow otherwise stopped running).
I am very new to running a relay and still getting up to speed - I imagine I will be for a while yet, so thanks everyone for your patience :)
Regards Zenaan