My tor node: https://metrics.torproject.org/rs.html#details/B1B10104EB72A1FBBF6687B05F191... Doesn't ever go up above 8800 or so. One thing I notice in Nyx is that my connections never go above about 2000 in and out connections. I have advertised bandwidth of just shy of a gigabit in my config. I understand now that the "advertised bandwidth" is calculated based on observed traffic through the node, which while more reliable and avoids abuse, seems to be counter productive to a degree. Ultimately what do I need to do to get more traffic through my node? Cause I have a 2Gbps fiber sitting here basically doing nothing so I was giving 1Gbps to tor :) Config -------------- RunAsDaemon 1 ControlPort 9051 ORPort 9001 ORPORT [2001:559:c290::fffb]:9001 RelayBandwidthRate 45 MB # Throttle traffic to 100KB/s (800Kbps) RelayBandwidthBurst 50 MB # But allow bursts up to 200KB/s (1600Kbps) DirPort 9030 # ExitPolicy reject *:* # no exits allowed ExitPolicy reject6 *:* Any suggestions appreciated. Matt Westfall President & CIO ECAN Solutions, Inc. Everything Computers and Networks 804.592.1672
Dear Matt, I see you started your relay 22 days ago. I started one (Aramis67) within a day of yours, so they both should still be getting adjusted. Have you read about the lifecycle of a new relay?: https://blog.torproject.org/comment/54651 As your new relay gets better known, I think your consensus number will go up, but while I have used RelayBandwidthRate to gradually cap my home relay to about 1TB a month, I don't think you can force a new relay up faster with bigger limits. I am running Aramis67 with no RelayBandwidthRate in the config: 15 ControlSocket /run/tor/control 16 ControlSocketsGroupWritable 1 17 CookieAuthentication 1 18 CookieAuthFile /run/tor/control.authcookie 19 CookieAuthFileGroupReadable 1 26 SOCKSPort 0 64 ControlPort 9051 91 ORPort 9001 92 ORPort [2a00:1dc0:caff:126::a5c9]:9001 187 ExitRelay 0 ----------------- My consensus weight is currently 5730, so a good deal behind 8400, though my current connections numbers are about 1800 and 2000 in and out. Much of the consensus weight seems to be affected by how close you are/good connection you have to the authentication servers. I expect that our relays will both generally go in the same direction. --Torix Sent with [ProtonMail](https://protonmail.com) Secure Email. ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐ On Friday, May 24, 2019 3:13 PM, Matt Westfall <mwestfall@ecansol.com> wrote:
My tor node: https://metrics.torproject.org/rs.html#details/B1B10104EB72A1FBBF6687B05F191...
Doesn't ever go up above 8800 or so.
One thing I notice in Nyx is that my connections never go above about 2000 in and out connections.
I have advertised bandwidth of just shy of a gigabit in my config. I understand now that the "advertised bandwidth" is calculated based on observed traffic through the node, which while more reliable and avoids abuse, seems to be counter productive to a degree.
Ultimately what do I need to do to get more traffic through my node? Cause I have a 2Gbps fiber sitting here basically doing nothing so I was giving 1Gbps to tor :)
Config -------------- RunAsDaemon 1 ControlPort 9051 ORPort 9001 ORPORT [2001:559:c290::fffb]:9001 RelayBandwidthRate 45 MB # Throttle traffic to 100KB/s (800Kbps) RelayBandwidthBurst 50 MB # But allow bursts up to 200KB/s (1600Kbps) DirPort 9030 # ExitPolicy reject *:* # no exits allowed ExitPolicy reject6 *:*
Any suggestions appreciated.
Matt Westfall President & CIO ECAN Solutions, Inc. Everything Computers and Networks 804.592.1672
Thanks for the feedback, however, I now have a new problem. I rebooted the server my TOR Node lives on and I've got a non-system disk error so I think my partition table got corrupted. If I can access the file system, is there any way to get what I need of the drive to retain my fingerprint? If I can't access the file system is there any way to get what I need from "somewhere" to retain my fingerprint? Matt Westfall President & CIO ECAN Solutions, Inc. Everything Computers and Networks 804.592.1672 http://ecansol.com On Sun, May 26, 2019 at 2:58 AM <torix@protonmail.com> wrote:
Dear Matt,
I see you started your relay 22 days ago. I started one (Aramis67) within a day of yours, so they both should still be getting adjusted. Have you read about the lifecycle of a new relay?: https://blog.torproject.org/comment/54651 As your new relay gets better known, I think your consensus number will go up, but while I have used RelayBandwidthRate to gradually cap my home relay to about 1TB a month, I don't think you can force a new relay up faster with bigger limits.
I am running Aramis67 with no RelayBandwidthRate in the config: 15 ControlSocket /run/tor/control 16 ControlSocketsGroupWritable 1 17 CookieAuthentication 1 18 CookieAuthFile /run/tor/control.authcookie 19 CookieAuthFileGroupReadable 1 26 SOCKSPort 0 64 ControlPort 9051 91 ORPort 9001 92 ORPort [2a00:1dc0:caff:126::a5c9]:9001 187 ExitRelay 0 ----------------- My consensus weight is currently 5730, so a good deal behind 8400, though my current connections numbers are about 1800 and 2000 in and out. Much of the consensus weight seems to be affected by how close you are/good connection you have to the authentication servers.
I expect that our relays will both generally go in the same direction.
--Torix
Sent with ProtonMail <https://protonmail.com> Secure Email.
‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐ On Friday, May 24, 2019 3:13 PM, Matt Westfall <mwestfall@ecansol.com> wrote:
My tor node: https://metrics.torproject.org/rs.html#details/B1B10104EB72A1FBBF6687B05F191...
Doesn't ever go up above 8800 or so.
One thing I notice in Nyx is that my connections never go above about 2000 in and out connections.
I have advertised bandwidth of just shy of a gigabit in my config. I understand now that the "advertised bandwidth" is calculated based on observed traffic through the node, which while more reliable and avoids abuse, seems to be counter productive to a degree.
Ultimately what do I need to do to get more traffic through my node? Cause I have a 2Gbps fiber sitting here basically doing nothing so I was giving 1Gbps to tor :)
Config -------------- RunAsDaemon 1 ControlPort 9051 ORPort 9001 ORPORT [2001:559:c290::fffb]:9001 RelayBandwidthRate 45 MB # Throttle traffic to 100KB/s (800Kbps) RelayBandwidthBurst 50 MB # But allow bursts up to 200KB/s (1600Kbps) DirPort 9030 # ExitPolicy reject *:* # no exits allowed ExitPolicy reject6 *:*
Any suggestions appreciated.
Matt Westfall President & CIO ECAN Solutions, Inc. Everything Computers and Networks 804.592.1672
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
On Thu, May 30, 2019 at 01:44:32PM -0400, Matt Westfall wrote:
Thanks for the feedback, however, I now have a new problem. I rebooted the server my TOR Node lives on and I've got a non-system disk error so I think my partition table got corrupted.
If I can access the file system, is there any way to get what I need of the drive to retain my fingerprint?
If I can't access the file system is there any way to get what I need from "somewhere" to retain my fingerprint?
You want to rescue the files in the "keys" directory in your DataDirectory. More details here: https://2019.www.torproject.org/docs/faq#UpgradeOrMove For example, on Debian or Ubuntu, you'll find them in /var/lib/tor/keys/ If the key files are gone, that fingerprint is gone from the world, and it's time to start a fresh one. :) Thanks! --Roger
Hello Matt, I recently started a new middle relay (torworld) as well, and also run a PT bridge on a separate network. This is just my knowledge, but tor relays are not always used to their maximum capacity, so your relay will receive more traffic some days and less on other days. Different tor relays are chosen for each user each time they connect, so varying amounts of traffic are to be expected. As long as your relay is correctly configured, you should not have much to worry about. --Keifer On Sat, May 25, 2019 at 4:48 PM Matt Westfall <mwestfall@ecansol.com> wrote:
My tor node: https://metrics.torproject.org/rs.html#details/B1B10104EB72A1FBBF6687B05F191...
Doesn't ever go up above 8800 or so.
One thing I notice in Nyx is that my connections never go above about 2000 in and out connections.
I have advertised bandwidth of just shy of a gigabit in my config. I understand now that the "advertised bandwidth" is calculated based on observed traffic through the node, which while more reliable and avoids abuse, seems to be counter productive to a degree.
Ultimately what do I need to do to get more traffic through my node? Cause I have a 2Gbps fiber sitting here basically doing nothing so I was giving 1Gbps to tor :)
Config -------------- RunAsDaemon 1 ControlPort 9051 ORPort 9001 ORPORT [2001:559:c290::fffb]:9001 RelayBandwidthRate 45 MB # Throttle traffic to 100KB/s (800Kbps) RelayBandwidthBurst 50 MB # But allow bursts up to 200KB/s (1600Kbps) DirPort 9030 # ExitPolicy reject *:* # no exits allowed ExitPolicy reject6 *:*
Any suggestions appreciated.
Matt Westfall President & CIO ECAN Solutions, Inc. Everything Computers and Networks 804.592.1672
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
Matt, if you only have 1 host, it may be more beneficial to create 2 relays on it (or more than 2 - if you have more than 1 IPv4 address available) using tor-instance-create. You could be hitting the limits of what a single CPU core can do. On Sun, May 26, 2019 at 4:07 PM Keifer Bly <keifer.bly@gmail.com> wrote:
Hello Matt, I recently started a new middle relay (torworld) as well, and also run a PT bridge on a separate network. This is just my knowledge, but tor relays are not always used to their maximum capacity, so your relay will receive more traffic some days and less on other days. Different tor relays are chosen for each user each time they connect, so varying amounts of traffic are to be expected. As long as your relay is correctly configured, you should not have much to worry about.
--Keifer
On Sat, May 25, 2019 at 4:48 PM Matt Westfall <mwestfall@ecansol.com> wrote:
My tor node: https://metrics.torproject.org/rs.html#details/B1B10104EB72A1FBBF6687B05F191...
Doesn't ever go up above 8800 or so.
One thing I notice in Nyx is that my connections never go above about 2000 in and out connections.
I have advertised bandwidth of just shy of a gigabit in my config. I understand now that the "advertised bandwidth" is calculated based on observed traffic through the node, which while more reliable and avoids abuse, seems to be counter productive to a degree.
Ultimately what do I need to do to get more traffic through my node? Cause I have a 2Gbps fiber sitting here basically doing nothing so I was giving 1Gbps to tor :)
Config -------------- RunAsDaemon 1 ControlPort 9051 ORPort 9001 ORPORT [2001:559:c290::fffb]:9001 RelayBandwidthRate 45 MB # Throttle traffic to 100KB/s (800Kbps) RelayBandwidthBurst 50 MB # But allow bursts up to 200KB/s (1600Kbps) DirPort 9030 # ExitPolicy reject *:* # no exits allowed ExitPolicy reject6 *:*
Any suggestions appreciated.
Matt Westfall President & CIO ECAN Solutions, Inc. Everything Computers and Networks 804.592.1672
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
Hi,
On 25 May 2019, at 01:13, Matt Westfall <mwestfall@ecansol.com> wrote:
My tor node: https://metrics.torproject.org/rs.html#details/B1B10104EB72A1FBBF6687B05F191...
Doesn't ever go up above 8800 or so.
One thing I notice in Nyx is that my connections never go above about 2000 in and out connections.
That's unusual, because there are about 7,000 relays in the network. How many simultaneous connections does your router support? (Lots of them claim to support unlimited connections, but only support a few hundred or a few thousand.)
I have advertised bandwidth of just shy of a gigabit in my config. I understand now that the "advertised bandwidth" is calculated based on observed traffic through the node, which while more reliable and avoids abuse, seems to be counter productive to a degree.
Ultimately what do I need to do to get more traffic through my node? Cause I have a 2Gbps fiber sitting here basically doing nothing so I was giving 1Gbps to tor :)
You could remove all the bandwidth limits, and put them back in when tor is using more than you want it to. (Tor tries to keep some extra bandwidth to deal with traffic spikes, so a 1 Gbps limit will get you around 300 kbps sustained traffic.) Here's a more detailed explanation, and some other things to try: https://trac.torproject.org/projects/tor/wiki/doc/MyRelayIsSlow T
Hello thanks for the comments, I might do that, remove the limits, because it's self limiting by the 1 Gbps network port, so it can't use more than that anyway. I'm using an Opnsense routing platform, and I've had more than 4,000 simultaneous connections just running torrents, lol. Igor, it doesn't appear to be a CPU bottleneck: https://puu.sh/DA4GR/2ef1b58e2e.png Am I able to run another tor instance just on different ports on the same IP? According to file descriptor limit I shouldn't be hitting a socket/file descriptor limit either. https://puu.sh/DA4Sh/6e27417c74.png I tried to run chutney tests to see what hardware supports but haven't quite figured out what the command line I should be using is. Any help with that would be appreciated. Thanks, Matt Westfall President & CIO ECAN Solutions, Inc. Everything Computers and Networks 804.592.1672 ------ Original Message ------ From: "teor" <teor@riseup.net> To: tor-relays@lists.torproject.org Sent: 5/30/2019 7:05:20 PM Subject: Re: [tor-relays] Relay Consensus Low
Hi,
On 25 May 2019, at 01:13, Matt Westfall <mwestfall@ecansol.com> wrote:
My tor node: https://metrics.torproject.org/rs.html#details/B1B10104EB72A1FBBF6687B05F191...
Doesn't ever go up above 8800 or so.
One thing I notice in Nyx is that my connections never go above about 2000 in and out connections.
That's unusual, because there are about 7,000 relays in the network. How many simultaneous connections does your router support? (Lots of them claim to support unlimited connections, but only support a few hundred or a few thousand.)
I have advertised bandwidth of just shy of a gigabit in my config. I understand now that the "advertised bandwidth" is calculated based on observed traffic through the node, which while more reliable and avoids abuse, seems to be counter productive to a degree.
Ultimately what do I need to do to get more traffic through my node? Cause I have a 2Gbps fiber sitting here basically doing nothing so I was giving 1Gbps to tor :)
You could remove all the bandwidth limits, and put them back in when tor is using more than you want it to. (Tor tries to keep some extra bandwidth to deal with traffic spikes, so a 1 Gbps limit will get you around 300 kbps sustained traffic.)
Here's a more detailed explanation, and some other things to try: https://trac.torproject.org/projects/tor/wiki/doc/MyRelayIsSlow
T
Hi,
On 1 Jun 2019, at 14:57, Matt Westfall <mwestfall@ecansol.com> wrote:
Hello thanks for the comments, I might do that, remove the limits, because it's self limiting by the 1 Gbps network port, so it can't use more than that anyway.
Following the instructions here: https://trac.torproject.org/projects/tor/wiki/doc/MyRelayIsSlow#TorNetworkLi... It looks like your relay is limited by its own observed bandwidth. (The observed bandwidth that your relay has seen itself using.) So increasing the RelayBandwidthRate would be a good idea. If your relay's observed bandwidth gets above 9 megabytes a second, your relay will be limited by the bandwidth authorities' measurements. (The median measurement for your relay is 8910 scaled kilobytes per second.) https://consensus-health.torproject.org/consensus-health-2019-06-02-04-00.ht... There might not be much you can do about this: Comcast has slow peering with a large number of internet networks. And it looks like 4/6 of tor's current bandwidth authorities are on those networks. This isn't something Tor can fix: we can only measure the bandwidth that Comcast is giving you. If Comcast has slow peering to US East and Europe, then clients using your relay will be slow.
I tried to run chutney tests to see what hardware supports but haven't quite figured out what the command line I should be using is.
Any help with that would be appreciated.
You're right, the README is more confusing than it needs to be. Try: ./chutney/tools/test-network.sh --data $[10*1024*1024] If a 10 MB transfer is too fast, try 100 MB. I opened this ticket for us to fix our documentation: https://trac.torproject.org/projects/tor/ticket/30720 T
Hey toer, I actually removed the Bandwidth Rates per another suggestion. Just sucks I can donate more to the TOR Network, but because other people abused the advertised bandwidth settings now it is what it is. https://puu.sh/DAw2V/aeb55530e8.png Also I guess the fact that most of the traffic across tor is http/https It's not ever going to "observe" a whole lot because it's quick small packets of data. I moved it to another IP and put it on 443 / 80 so maybe that will help cause firewalls and such. It's also directly on a public wan IP now, so firewall/router complications. Config in Nyx: https://puu.sh/DAwXD/39269fba87.png Chutney Results - https://puu.sh/DAwYD/131f6f1959.png Ran a 30 MB Test 10 was fast and 100 kept crashing 2nd Run: https://puu.sh/DAx0v/4c73a8b661.png So looks like my hardware can only handle about 100Mbps Full Duplex, but that's still way more than 9 :-D Guess we'll see what happens. I didn't see if anyone answered if I need a separate IP or if I can create another tor instance on different ports but on the same IP, to increase the load I'm handling. Thanks, Matt Westfall President & CIO ECAN Solutions, Inc. Everything Computers and Networks 804.592.1672 ------ Original Message ------ From: "teor" <teor@riseup.net> To: tor-relays@lists.torproject.org Sent: 6/2/2019 1:45:35 AM Subject: Re: [tor-relays] Relay Consensus Low
Hi,
On 1 Jun 2019, at 14:57, Matt Westfall <mwestfall@ecansol.com> wrote:
Hello thanks for the comments, I might do that, remove the limits, because it's self limiting by the 1 Gbps network port, so it can't use more than that anyway.
Following the instructions here: https://trac.torproject.org/projects/tor/wiki/doc/MyRelayIsSlow#TorNetworkLi...
It looks like your relay is limited by its own observed bandwidth. (The observed bandwidth that your relay has seen itself using.)
So increasing the RelayBandwidthRate would be a good idea.
If your relay's observed bandwidth gets above 9 megabytes a second, your relay will be limited by the bandwidth authorities' measurements. (The median measurement for your relay is 8910 scaled kilobytes per second.)
https://consensus-health.torproject.org/consensus-health-2019-06-02-04-00.ht...
There might not be much you can do about this: Comcast has slow peering with a large number of internet networks. And it looks like 4/6 of tor's current bandwidth authorities are on those networks.
This isn't something Tor can fix: we can only measure the bandwidth that Comcast is giving you. If Comcast has slow peering to US East and Europe, then clients using your relay will be slow.
I tried to run chutney tests to see what hardware supports but haven't quite figured out what the command line I should be using is.
Any help with that would be appreciated.
You're right, the README is more confusing than it needs to be.
Try: ./chutney/tools/test-network.sh --data $[10*1024*1024]
If a 10 MB transfer is too fast, try 100 MB.
I opened this ticket for us to fix our documentation: https://trac.torproject.org/projects/tor/ticket/30720
T
Hi,
On 2 Jun 2019, at 16:57, Matt Westfall <mwestfall@ecansol.com> wrote:
Hey toer, I actually removed the Bandwidth Rates per another suggestion.
You might need to wait a week or two for the new setting to increase your bandwidth. It takes a few days for the bandwidth authorities to measure the whole network.
Just sucks I can donate more to the TOR Network, but because other people abused the advertised bandwidth settings now it is what it is.
I'm sorry, but Comcast's peering is the problem here, not Tor. Tor clients choose random relays, regardless of their location. So we need to measure how well your relay connects to the rest of the world. Lots of relay operators expect Tor to use all their bandwidth. But Tor is low-latency, so it should never use all a relay's bandwidth. (10% is ideal, 30% is typical, any more than that causes lots of latency for clients.) There's a detailed explanation on this wiki page: https://trac.torproject.org/projects/tor/wiki/doc/MyRelayIsSlow#WhyRelayLoad...
Also I guess the fact that most of the traffic across tor is http/https It's not ever going to "observe" a whole lot because it's quick small packets of data.
Tor Browser regularly downloads 100MB+ updates over Tor. And people use Tor for other bulk downloads. And remember: clients choose relays at random, so all relays see a similar mix of small and large downloads.
I moved it to another IP and put it on 443 / 80 so maybe that will help cause firewalls and such. It's also directly on a public wan IP now, so firewall/router complications.
If you keep your relay stable for a few weeks, it will probably get the Guard flag again. Relays with the Guard flag get more traffic.
I didn't see if anyone answered if I need a separate IP or if I can create another tor instance on different ports but on the same IP, to increase the load I'm handling.
You can create 2 relays per IPv4 address:
On 27 May 2019, at 11:54, Igor Mitrofanov <igor.n.mitrofanov@gmail.com> wrote:
Matt, if you only have 1 host, it may be more beneficial to create 2 relays on it (or more than 2 - if you have more than 1 IPv4 address available) using tor-instance-create. You could be hitting the limits of what a single CPU core can do.
Even if your relay isn't used much for client traffic, it still works well as a directory mirror for relay and onion service information. If you'd like Tor to use more of your traffic, and your relay will be on the same address and port for the next 2 years, you can sign up as a fallback directory mirror: https://lists.torproject.org/pipermail/tor-relays/2019-May/017321.html Since you just changed your relay port, your relay won't be included in the June 2019 list. But we'll do another list in 6-12 months. T
participants (6)
-
Igor Mitrofanov -
Keifer Bly -
Matt Westfall -
Roger Dingledine -
teor -
torix＠protonmail.com