- tor-relays - lists.torproject.org

Relay requirements
by Sydney 10 Mar '23

10 Mar '23

Newbie here. No network experience but already running 2 TOR instances: 1 TOR service + 1 bridge. I would like to "upgrade" to TOR relays but have a few questions relating to hardware needs. The [TOR-friendly] ISP I'm looking at have these 4 offers available: [1] 1 core @ 3.1GHz 1 GB RAM 10 TB/mo @ 5 Gbps BW [2] 1 core @ 3.1GHz 2 GB RAM10 TB/mo @ 5 Gbps BW [3] 2 core @ 3.1GHz4 GB RAM20 TB/mo @ 5 Gbps BW [4] 3 core @ 3.1GHz 12 GB RAM30 TB/mo @ 5 Gbps BW From [relay requirements](https://community.torproject.org/relay/relays-requirements/), my understanding is that [1] is what is expected for a guard or middle relay and [2] for an exit relay. My preferred choice at the moment is [3] where I would like to have 3 relays (one of each type; 3.5 GB RAM needed). Each relay would be limited to, say, 6.6 TB/mo BW. This would allow me to experiment and get used to the network and its requirements. My questions is about future increase. If I want to increase my offer to the TOR network I could: - Negotiate with my ISP to increase my BW to 30 TB/mo while still running on [3] and increase each BW limit to 10 TB/mo; - Move my 3 relays on [4] and increase each BW limit to 10 TB/mo; - Move my 3 relays on [4], add 2 IP addresses and 6 relays (2 of each type; 10.5 GB RAM needed total), and limit each relay to 3.3 TB/mo. I guess my fundamental question is what is the advantage of running multiple relays of the same type, on the same server? I see some operators running dozens of them, all in the same country, same ISP. Why not just a single relay running with a large capacity? Also, is there a requirement for the number of relays per core? (Maybe this is the answer to my question.) I know my bridge is currently keeping one core of my 2-core server constantly under load. Thank in advance.

3 3

Draft workshop submission "Running a high-performance pluggable transports Tor bridge"
by David Fifield 07 Mar '23

07 Mar '23

Linus Nordberg and I have been working together to run the main Snowflake bridge since April 2022. We are preparing a short paper (4 pages) for the FOCI workshop (https://foci.community/) on the special procedures required to operate a bridge that gets the large volume of traffic that a Snowflake bridge does. This is a draft of the submission: https://www.bamsoftware.com/papers/pt-bridge-hiperf/pt-bridge-hiperf.202303… https://www.bamsoftware.com/papers/pt-bridge-hiperf/pt-bridge-hiperf.202303… I am posting the draft here to see if anyone has feedback or comments. If you do, our submission deadline is 2023-03-15. The main ideas were worked out in a thread on this very list: https://forum.torproject.net/t/tor-relays-how-to-reduce-tor-cpu-load-on-a-s… The core problem is that a single tor process is limited to a single CPU core. The main fix is to run multiple tor processes, and synchronize their identity and onion keys externally. It has been essential to enable scaling in the Snowflake bridge; but the same trick may be useful (minus the pluggable transport part) for other relays that process a much larger amount of traffic than average, such as large exit relays. (Though it's not clear that one exit with 2× the capacity is really any different than 2 exits in the same family.) This is also somewhat related to the current discussion about increasing the limit on the number of relays per IP address, which is a concession to the limited scalability of one tor process: https://bugs.torproject.org/tpo/core/tor/40744 > While this is great in general, it also adds significant complexity > for Tor operators since Tor does not scale well on multi-core CPU's. > In order to effectively use a modern server with a large amount of > threads, you would need to run many Tor relays.

1 0

D5A3882CBDBE4CAD2F9DDA2AB80FE761BEDC3F11 is spoofing my contact info
by lists＠for-privacy.net 06 Mar '23

06 Mar '23

This is _not_ my relay: https://metrics.torproject.org/rs.html#details/D5A3882CBDBE4CAD2F9DDA2AB80F… https://nusenu.github.io/OrNetStats/w/relay/D5A3882CBDBE4CAD2F9DDA2AB80FE76… -- ╰_╯ Ciao Marco! Debian GNU/Linux It's free software and it gives you freedom!

3 2

Storing ed25519_master_id_secret_key[_encrypted] on a smartcard?
by telekobold 04 Mar '23

04 Mar '23

Dear fellow relay operators, currently, I'm operating a Tor relay (Middle/Guard) and a Tor Bridge. Offline keys [1],[2] are a good way to secure a Tor relay, but I'm wondering if there is a standard way or something like a hacking guide how to store your ed25519_master_id_secret_key[_encrypted] on a smartcard or hardware token like a Nitrokey or Yubikey? This would even be more secure than storing it on a "normal" USB device. Unfortunately I have not found much about this on the internet. Kind regards telekobold [1] https://support.torproject.org/relay-operators/offline-ed25519/ [2] https://gitlab.torproject.org/legacy/trac/-/wikis/doc/TorRelaySecurity/Offl…

1 1

RFC: does a private exit would work?
by Toralf Förster 04 Mar '23

04 Mar '23

tl;dr; restricted access + usage of an exit longer: An exit is sooner or later abused. A reduced exit policy does not prevent that. What about setup a tor exit relay with 'PublishServerDescriptor = 0' ? Having an access line like for bridges would restrict the access. An alternative could be a port knockig + iptables solution. Objections and comments are welcome. -- Toralf

5 5

Preferred range for ephemeral ports
by binarynoise 03 Mar '23

03 Mar '23

Hello everyone, I just set up a snowflake proxy on my pi at home. Unfortunately I can only forward 255 ports to my pi (at once, or is there a way to trick my Fritz!box into allowing more?). Which is the preferred range for -ephemeral-ports-range when it needs to be restricted to allow the most users to connect? Is 60000:60250 ok? Greetings binarynoise

2 1

removing tor versions from "recommended versions" before they reach EOL
by nusenu 01 Mar '23

01 Mar '23

Hi, the tor 0.4.5.x end of live versions are still on the recommended versions list: https://consensus-health.torproject.org/#recommendedversions > consensus client-versions 0.4.5.6, 0.4.5.7, 0.4.5.8, 0.4.5.9, 0.4.5.10, 0.4.5.11, 0.4.5.12, 0.4.5.14, 0.4.5.15, 0.4.5.16, 0.4.7.7, 0.4.7.8, 0.4.7.10, 0.4.7.11, 0.4.7.12, 0.4.7.13 > server-versions 0.4.5.6, 0.4.5.7, 0.4.5.8, 0.4.5.9, 0.4.5.10, 0.4.5.11, 0.4.5.12, 0.4.5.14, 0.4.5.15, 0.4.5.16, 0.4.7.7, 0.4.7.8, 0.4.7.10, 0.4.7.11, 0.4.7.12, 0.4.7.13 it would probably be better to stop recommending tor versions a month before they reach their eol date so relay operators get to see a log entry and have some time to react. kind regards, nusenu -- https://nusenu.github.io

2 1

relay configuration not allowing optimal performance
by Relayer 26 Feb '23

26 Feb '23

I'm running a tor relay on some older hardware that I didn't want to discard when I could still put it so good use. Some details of the box are: -- CPU: Intel(R) Core(TM)2 Duo CPU P8600 @ 2.40GHz -- RAM: 4GB -- ARCH: x86_64 -- HDD: 250GB -- OS: Ubuntu 22.04.1 I originally configured a single Tor instance IPv4 to run as a relay only (not as an exit, nor hosting a hidden service). I am also using the iptables rules from https://github.com/Enkidu-6/tor-ddos to minimize DDOS overhead (please advise if there are alternatives or additions to this). My original config seemed ok until I started seeing my CPU and RAM maxing out consistently so I throttled back with the following in my torrc: RelayBandwidthRate 100 KB # Throttle traffic to 100KB/s (800Kbps) RelayBandwidthBurst 200 KB # But allow bursts up to 200KB/s (1600Kbps) MaxAdvertisedBandwidth 1 MB My RAM usage now is only about 50% or marginally less of my total available. Here's how the metrics look lately: https://metrics.torproject.org/rs.html#details/38939B45237BA84941C74836349C… As you can see, the throughput rated dropped in half (that's when the graph drops on 2023-02-09). However, the volume continued to decline. Additionally, I'm unclear why my Middle Probability and Consensus Weight have both dropped to near 0%. Are those, in fact, where I want them? I'm monitoring with nyx and see I get some traffic through with no apparent errors or warnings. I am NOT seeing the CPU spikes any longer but I don't think I'm giving the most with my hardware. Questions: 1.) Is my tor service now misconfigured and not utilizing my hardware as best it could? 2.) Should my Consensus Weight and/or Middle Probability be higher? 3.) Should I consider running two tor instances? Nyx log snippet: 07:59:32 [NOTICE] Heartbeat: DoS mitigation since startup: 7 circuits killed with too many cells, 591 circuits rejected, 2 marked addresses, 0 marked addresses for max queue, 0 same address concurrent │ connections rejected, 0 connections rejected, 0 single hop clients refused, 19166 INTRODUCE2 rejected. [1 duplicate hidden] │ 07:59:32 [NOTICE] Since startup we initiated 0 and received 0 v1 connections; initiated 0 and received 0 v2 connections; initiated 0 and received 0 v3 connections; initiated 0 and received 57982 v4 │ connections; initiated 116266 and received 356623 v5 connections. │ 07:59:32 [NOTICE] Circuit handshake stats since last time: 3/3 TAP, 44849/44849 NTor. [1 duplicate hidden] │ 07:59:32 [NOTICE] While not bootstrapping, fetched this many bytes: 194128391 (server descriptor fetch); 7140 (server descriptor upload); 17539422 (consensus network-status fetch); 1794 (authority cert │ fetch); 2111765 (microdescriptor fetch) │ 07:59:32 [NOTICE] Heartbeat: Tor's uptime is 10 days 23:58 hours, with 179 circuits open. I've sent 34.83 GB and received 35.63 GB. I've received 444762 connections on IPv4 and 0 on IPv6. I've made │ 254336 connections with IPv4 and 0 with IPv6. [1 duplicate hidden] │ 01:59:32 [NOTICE] Since startup we initiated 0 and received 0 v1 connections; initiated 0 and received 0 v2 connections; initiated 0 and received 0 v3 connections; initiated 0 and received 56651 v4 │ connections; initiated 114326 and received 347071 v5 connections. │ 01:59:32 [NOTICE] While not bootstrapping, fetched this many bytes: 189431170 (server descriptor fetch); 7140 (server descriptor upload); 17131743 (consensus network-status fetch); 1794 (authority cert │ fetch); 2068377 (microdescriptor fetch) Thanks. Relayer1974

2 1

Re: [tor-relays] Frantech (was Re: Confusing bridge signs)
by Jeff Teitel 26 Feb '23

26 Feb '23

On Fri, Feb 24, 2023 at 12:10 AM Marco wrote: > Yes, Frantech should actually be avoided. But in Miami there are few Tor > relays. A SLICE 512 for $2.00/m or $20.00/y is sufficient for a bridge. > https://buyvm.net/kvm-dedicated-server-slices/ Why should Frantech be avoided? Jeff Teitel

2 1

Questions about Tor consensus weight & swag
by shruub 19 Feb '23

19 Feb '23

Hello, so my relay regulary gets overloaded, for what I can only assume is an hardware issue, (which I can't upgrade rn) since I already applied the tor anti ddos scripts. It seems to be able to recover, however with a lower Consensus Weight. I also (stupidly) tried to have a cron restarting my tor daemon daily which also resulted in the latter. So I wonder, if there is any way to have a relay run more stable and I suppose with a somewhat higher consensus weight (I can only asssume making some further changes in advertising bandwidth etc. My second question(s) is/are concerning the tor swag (I hope that's allowed to ask here). Firstly, what is the actual Speed requirement? In the tor ecosystem, every unit is MBs, and only on the swag site its KBs. But if my calculations are correct, 500KBs = 500000Bs = 0.5MBs which doesn't really make sence, imo(but I probably misscalculated somewhere). Secondly, does running mean it's uptime (aka Last Restarted) or the Time it was first seen? Sorry for the wall of text for not much information, and thanks a lot in advance for your replies! Best, shruub

4 5