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

Re: The speed of my relay are decreasing and my rely now is out of the Tor statistics



     On Sun, 14 Feb 2010 22:07:01 -0500 andrew@xxxxxxxxxxxxxx wrote:
>On Sun, Feb 14, 2010 at 09:43:34AM +0100, ran6oony7r9deku5@xxxxxxxxxxxxxx wrote 0.4K bytes in 14 lines about:
>: >> This happens every night, but after that my Tor node is not used
>: >> anymore until I restart my Tor server (TorDwarf).
>: > 
>: > Are you sure this is true?
>: 
>: Yes - without restart there ist no (normal) traffic after > 2 hours.
>
>Unless there is some fine new bug, this is not true.  I've run a relay on a

     I think it may well be true *now*.

>dynamic IP that changes roughly weekly, and it saw traffic fairly
>consistently throughout the week with a slight dip for the 1-2 hours after
>the IP changed.
>
     Have you tried it since the switch to consensus documents?  As I noted
in my most recent posting to OR-TALK, a newly initialized relay does not
appear in the next consensus document issued, but rather in the one that
follows it.  That right there accounts for up to two hours' delay in clients
finding out about the new relay.  Now that new consensus documents are
usually not fetched by clients every hour anymore, but rather on a seemingly
random interval ranging from ~30 min. to ~2 hrs. 30 min., a client might
very well not know about a new relay for over four hours.
     Changing the IP address causes reinitialization, including reachability
and data rate capacity testing.  Given the right timing of the IP address
change, the authorities' own reachability testing of relays in the directory
might fall at a time when the address being tested had changed, but before
the relay in question had discovered that its address had changed,
reinitialized itself, and published its new descriptor, causing the
authorities to decide that the relay was no longer reachable and to drop the
relay from the next consensus.  That, in turn, would mean that clients would
not know that a new descriptor were available and thus would not download it
from a directory server.  It would not be surprising to me at all to find
cases where some fraction of the client population did not know about a
relay becoming operational at a new IP address for well over four hours.
Given that the most recent consensus document on my system at this moment
claims that TorDwarf's capacity is only "24", that relay will only be chosen
once in a while for a circuit and therefore will probably show very little
traffic for the first two or three hours after reinitialization at a new IP
address.


                                  Scott Bennett, Comm. ASMELG, CFIAG
**********************************************************************
* Internet:       bennett at cs.niu.edu                              *
*--------------------------------------------------------------------*
* "A well regulated and disciplined militia, is at all times a good  *
* objection to the introduction of that bane of all free governments *
* -- a standing army."                                               *
*    -- Gov. John Hancock, New York Journal, 28 January 1790         *
**********************************************************************