[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: TOR dying



On Fri, Apr 15, 2005 at 10:13:19AM -0400, Jonathan D. Proulx wrote:
> 
> My TOR server (0.1.0.3-rc) crashed at some point yesterday.  It had
> been running fine for a few days so I don't think too much of it
> untill I came in this morning and found it dead again.
> 
> The relevent log section seems to be the second of these (others given
> for context):
> 
> Apr 15 08:47:14.853 [warn] rend_mid_rendezvous(): Rejecting RENDEZVOUS1 cell wit
> h unrecognized rendezvous cookie 38D49D1D
> Apr 15 09:12:44.957 [err] do_main_loop(): libevent poll failed: Invalid argument
>  [22]
> Apr 15 09:31:01.705 [notice] Tor 0.1.0.3-rc opening log file.
> 
> I'm running Debian "unstable" with TOR from the noreply.org
> experimental repository (0.1.0.3-rc-200504080730-1)

Dan,

Thanks for your report.  I have also been running into the same problem,
and I submitted it as a bug report:

http://bugs.noreply.org/flyspray/index.php?do=details&id=117

My suspicion is that the problem is related to the fact that both rodos
(my server with the problem) and spoon (presumably your server with the
problem) are among the very fastest, with allocation of the most file
descriptors.

I also suspect that you and I are both running the ancient version of
libevent that exists in the form of a Debian package.  Perhaps if we
could pester Niels Provos nicely, he might be inclined to give us
something that does not have the rare bug exposed by fast Tor servers.
;)  Anyway, just a thought.

Please add more details on the Flyspray page if you have more to give.

Thanks,

Geoff

Attachment: signature.asc
Description: Digital signature