February 2015 - tor-relays - lists.torproject.org

Thoughts on new relay
by Niklas Femerstrand 11 Feb '15

11 Feb '15

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Moin, Our relay is online as of Jan 23 and recently I've been thinking about pivoting it into a bridge instead. The relay lives in Cambodia and may naturally be a bit off from the rest of the network (latency). Consensus weight is steadily increasing despite various initial downtime incidents, which you can see in Atlas graphs. (We've had some network upgrades and during a separate event even had a monkey yanking the eth from the nic.) Currently we advertise 435.66 KB/s and with our current ~9d 10h uptime we have downloaded 3.6 GB resp uploaded 3.7 GB. Theoretically our tubes fit 10 Mbps download and 30 Mbps upload. The question is whether the relay is worth keeping with its current setup or if we can contribute more by re-configuring into a bridge. Unfortunately it is very risky for us to allow exit traffic, but we <3 any cipher packet. Please advice! https://atlas.torproject.org/#details/94F806110B23E727AE8BAD74DD95B0EE0B7A8… Regards, King Kong -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBCAAGBQJU2gGzAAoJEHxAR3sGx+rLTzEP/05TJK3vzfPe3kUjUfg0G/wm vlc1vZLshAvK/pE1mf2DqDSgnuxR+QHHJI8htJOaU5VJVMlnEy4nsF0uE76It1EW OLUo7+LU452Sx7XjyVYMKOu6OJ9kD1Cg6OkNT2HEIOR+57x7vNJTYr7oF0uNu6dt UzVFJKkM3Y1ca4GsMS2M/oXFw2jGrrvmx4DLHu2OSDwNumy3ndOETnZscVfErAMK A/K4JI0Ws6pInBRciEHdak9ZBkR/f9jYGPbUXOSwnJN6XIsawijTIbrRLfYiUcro 0wj6mZYqXgKRDOvP57UNkf5UaolADQb6ssiLkhhAGu+DHnTt5V8g2vkqtY3wU02N 823X7o3oNZYq5yZFTbexH8Al+ys9KvKlu5q0YRYvbBwvI/yGHyr/gEI7ufV+9tZ6 z9Mo2hCBsaM0vrikBzejhjSw5qc4HcwMKErlC3cHQHIbszryu1O0WY1G22pwYgss mAGDr6OcdeIdMyNcJlpyH8+jToRSJrkAwDSPdQz+IVMq785x0GnOo2uV2/7NTr2J +LVUjSjAenRhPX+TrAFFujNYXZf5uD/EUfsGyUyCdBhJ5PWsLjYQ/EYD6PToDX8P BdocR3XfxZS7nD5PYRDfyWNeckCL6VPrnZrgMGIYBEXAn7jhX608vzmZV7IU7mzv 3NrnDWpllK4M42hT7fXs =hV51 -----END PGP SIGNATURE-----

3 3

Is Bridges supposed to get a gurard flag?.
by Sasikantha babu 11 Feb '15

11 Feb '15

Hello All, I've been running a non-exit relay at home and a bridge on amazon cloud since dec-2014. And today I had noticed on atlas my bridge has got a guard flag and from the logs in the last 18 hours it had sent 4.73 GB and received 4.89 GB (finally I'm happy to see bandwidth being used on bridge). I've the following questions 1. Is a bridge supposed to get a guard flag ?. 2. Is there a difference between a guard flag of a relay and a guard flag of a bridge?. Thanks

2 1

question and concering about probability of exitnode
by tor＠viisauksena.de 10 Feb '15

10 Feb '15

hi, just want to note, i am proud to have an german exit node with 11Mbs .. but also concerned about this raising exit node probability, actually it reads for me like 1 of 400 tor-users is going through this exit (by recalculating a exit probability of 0.25 %) - this concerns me because it should not be so easy (running just some days on a vserver) and i don't know how high this will raise ? i mean, its easy for me to guess a nsa/gchq/bnd bad-ass-club can get some hundred vservers running, and get tor exclusively exits .. and guard (which was much more easy, i lost it because i quickly run out of maximum limit for a month, so i cut it, but the 1 day downtime lost me the guard flag) just my opinion .. and if you can show me some more information over the numbers of exits and guards and so on, i would be happy, cant believe there are so less many exit nodes out there ... thx, Jens

2 2

Minimizing service interruption when reloading torrc?
by Elliott Jin 09 Feb '15

09 Feb '15

Hi all, I'd like to reload torrc; I think I can do this by either restarting tor or by sending a SIGHUP to the tor process, but I'm not sure how much of a service interruption each of these approaches will cause. What's the best way to reload torrc (i.e. how can I do this in a way that minimizes the service interruption)? Thanks! -Elliott -- Anonymous Feedback <http://www.admonymous.com/eyjin>

3 3

Relay operators: help improve this hardening document?
by Nick Mathewson 07 Feb '15

07 Feb '15

Hi, all! There's a project going on to try to add instructions for hardening a Tor relay for security: https://trac.torproject.org/projects/tor/ticket/13703 The idea is that Tor could ship with some basic recommendations, and links to places to find more advice? Recently, "mmcc" has uploaded a new draft. Do we think this is better than nothing and worth shipping with Tor, or does it need big changes? If possible, please write comments on the trac ticket above: it will help keep all the discussion in one place. best wishes, -- Nick

7 6

Changes in network traffic pattern
by Hu Man 07 Feb '15

07 Feb '15

Hi All I have been running a tor relay for about a year and according to my munin graph It normally receives, on average, just under 2,000 incoming tcp connections on port 443 every 5 minutes. In the last few days that figure has increased to about 10,000 and spiked to about 19,000 incoming requests every 5 minutes. First thought was DDOS but traffic is not high enough to cause any problems. I did some digging and in a 5 minute period received the following requests to the port tor is listening on (number of requests and source ip address) 2722 SRC=107.167.22.79 1355 SRC=107.167.22.90 1334 SRC=104.37.244.131 1237 SRC=213.251.185.14 604 SRC=188.247.130.32 13 DST=178.200.216.58 7 SRC=92.63.110.232 6 SRC=5.196.8.208 6 SRC=200.76.82.231 6 DST=93.158.248.243 This is only the top 10 source ip addresses. I had a look and none of the top few seem to be tor relays. Just wondering if others are seeing a large number of requests from the above ip addresses or if it's just me. If it is just me then I can easily just block these ip addresses.

3 2

Avoiding using my own Exit
by Chris Tilt 07 Feb '15

07 Feb '15

Hello, When running a relay that is configured as an Exit, do you run the risk of your own TOR browsing sessions (initiated from the same private network) using the relay and effectively short-circuiting the location randomization? Note that I don't have a relay up yet, so my question may be lacking clarity. Cheers, Chris

2 1

TOR: Inbound, Outbound, Exit connections
by Ralph Bolliger 06 Feb '15

06 Feb '15

*Good evening ladies and gentleman* I'm running a TOR Exit for a few days now ( https://globe.torproject.org/#/relay/06BA80D9E1143CFAD835442142A3FA5A1E4FD9…) I'm also using TOR ARM in order to monitor TOR's performance, log messages and connections. When I have a look at the connections page on TOR ARM I read about «Inbound», «Outbound» or «Exit» connections. I searched the web already. But I wasn't able to find a site that explains in simple words what's the difference between «Inbound», «Outbound» or «Exit» connections. Is there anybody who is able to tell me what «Inbound», «Outbound» or «Exit» connections are? Greetings from Switzerland Information Architect

2 1

Making your relay submit hidden service statistics
by George Kadianakis 04 Feb '15

04 Feb '15

Hello people, some days ago we sent a mail [1] asking for feedback on our patch that collects hidden service statistics from Tor relays. Specifically, we are trying to answer the questions: "Approximately how many hidden services are there?" and "Approximately how much traffic of the Tor network is going to hidden services?". For more information see my previous email and the relevant proposals. Our code [2] got merged in upstream Tor some days ago and now we are looking for more volunteers to enable it on their relays. We will collect these statistics during the next [days] weeks, and in a month we plan to publish a short report with our findings. To help us you will need to use the very latest version of Tor and also slightly edit your torrc. Here are some instructions for Debian/Ubuntu [3]: ----- Let's start by slightly editing your regular Tor torrc to enable hidden service statistics. Please append the folowing to your regular torrc (not including quotes): """ HiddenServiceStatistics 1 """ Now, we need to install a bleeding edge version of Tor that can do hidden service statistics. You can either use the git master [3], or use Tor's APT repositories. The following instructions help you with the latter approach. Please start by following "Option two" of https://www.torproject.org/docs/debian so that you add the proper APT repositories and install tor and the deb.torproject.org-keyring package. Next, add the following line to your sources.lst: deb http://deb.torproject.org/torproject.org tor-nightly-master-wheezy main (you might need to replace 'wheezy' with your distribution, like you did in the previous step) Now, do # apt-get update # apt-get install tor which should upgrade your Tor to the latest nightly version., Your system Tor should now be up-to-date and able to collect hidden service statistics! Just start Tor as you normally do, using the torrc we modified above. For example you could simply do: # tor -f torrc but feel free to use your init script if you prefer that. Now just leave your relay do its thing, and keep an eye to this mailing list around mid-January 2015 for the results of our research. --- FWIW, if you are not using Debian, you can use the master branch of the official git repository: https://git.torproject.org/tor.git You will also need to edit your torrc in the same manner as the Debian instructions. Feel free to reply with any questions you have! Thanks! George Kadianakis Karsten Loesing Aaron Johnson David Goulet [1]: https://lists.torproject.org/pipermail/tor-dev/2014-December/007968.html [2]: https://trac.torproject.org/projects/tor/ticket/13192 [3]: for example see http://tor.stackexchange.com/questions/75/how-can-i-install-tor-from-the-so…

1 1

Long running relay keeping low bandwidth utilization
by mail＠dotjhw.pw 04 Feb '15

04 Feb '15

Dear list, I am wondering, wether it is still within what is considered a normal lifecycle, if the relay I am running is utilizing significantly less than the available bandwidth? Some background info: I set the relay up a few months ago and experienced higher utilization after a few days. around the end of 2014, I experienced hardware failure with the relay's machine and it went offline on December 21. I just managed to bring it back online on January 13. Since then, even with an available bandwidth in the range of 5-10Mbit/s, the actually used bandwidth is averaging at 8-9kbit/s, with no increase in sight. As far the bandwidths and weights graphs are concerned, it seems like the relay is not weighted as high anymore (last known consensus weight before the crash: 2.22e-5. After the crash and rebuild: 2.01e-6). Is there anything I can do to improve the relay? Best regards, .jhw

2 2