- tor-relays - lists.torproject.org

Re: [tor-relays] Circuit creation "storms" overwhelming Raspberry Pi?
by temp5＠tormail.org 06 Jun '13

06 Jun '13

I've been seeing these storms as well on my relay. I average something like 100 connections for weeks and weeks per the tor logs, but then suddenly it will jump into the thousands and I'll see the "Failed to hand off onionskin." and "Your computer is too slow to handle this many circuit creation requests!" messages. I wonder if it's some type of DDOS too. I thought about using this method (http://www.debian-administration.org/articles/187) on the relay and dir ports, but I'm not sure what sort of limits to set. Like does 1 Tor circuit = 1 iptables connection? Or if a user hits a webpage with 100 ads on it, maybe it would be 1 Tor circuit = 100 iptables connections? That's about as far as I got. I didn't want to break things by trying to fix another. > I did a lot of tuning on the Raspberry Pi and it's now much, much more > stable as a Tor relay, but just now I had another "circuit creation > storm." Interestingly, the Pi remained up, and my *router* crashed. > I've also seen huge bursts of circuit creation on a relay I run on a > VPS, but as it's a much more powerful box it rarely complains (and thus > I rarely notice). > > This many circuits and outbound connections is highly unusual for the > small relay I'm running on the Pi. And this behavior definitely occurs > in bursts. Is this an outbound DDOS or an attack on Tor itself? If the > former (or maybe the latter), is there some way I could perhaps use > iptables to temporarily "clamp" the ability to open TCP connections when > Tor (or anything on the Pi) opens a number over some threshold in some > short period of time? > > Here's log output (via 'arm') from the relay after my router crashed > twice, I went to the admin panel and noted hundreds of outbound > connections from my Tor box. Time is America/Los_Angeles. > > │ 13:55:00 [ARM_NOTICE] Relay unresponsive (last heartbeat: Sat May > 4 13:54:14 2013) > │ 13:52:25 [WARN] Your computer is too slow to handle this many > circuit > creation > │ requests! Please consider using the MaxAdvertisedBandwidth > config > option or choosing > │ a more restricted exit policy. [404 similar message(s) > suppressed > in last 60 seconds] > │ 13:51:07 [WARN] Your computer is too slow to handle this many > circuit > creation > │ requests! Please consider using the MaxAdvertisedBandwidth > config > option or choosing > │ a more restricted exit policy. [75 similar message(s) > suppressed in > last 60 seconds] > │ 13:50:52 [WARN] Your computer is too slow to handle this many > circuit > creation > │ requests! Please consider using the MaxAdvertisedBandwidth > config > option or choosing > │ a more restricted exit policy. [601 similar message(s) > suppressed > in last 60 seconds] > │ 13:48:39 [WARN] Your computer is too slow to handle this many > circuit > creation > │ requests! Please consider using the MaxAdvertisedBandwidth > config > option or choosing > │ a more restricted exit policy. [99 similar message(s) > suppressed in > last 60 seconds] > │ 13:47:34 [WARN] Your computer is too slow to handle this many > circuit > creation > │ requests! Please consider using the MaxAdvertisedBandwidth > config > option or choosing > │ a more restricted exit policy. [22 similar message(s) > suppressed in > last 60 seconds] > │ 13:46:17 [WARN] Your computer is too slow to handle this many > circuit > creation > │ requests! Please consider using the MaxAdvertisedBandwidth > config > option or choosing > │ a more restricted exit policy. [253 similar message(s) > suppressed > in last 60 seconds] > │ 13:43:47 [WARN] Your computer is too slow to handle this many > circuit > creation > │ requests! Please consider using the MaxAdvertisedBandwidth > config > option or choosing > │ a more restricted exit policy. [1396 similar message(s) > suppressed > in last 60 > │ seconds] > │ 13:42:48 [WARN] Your computer is too slow to handle this many > circuit > creation > │ requests! Please consider using the MaxAdvertisedBandwidth > config > option or choosing > │ a more restricted exit policy. [16 similar message(s) > suppressed in > last 60 seconds] > > Here's how it crashed my router (blowing ip_conntrack limits is > sufficient only to mess up many of my TCP connections, but eventually > the router runs out of memory and starts killing processes): > > May 4 13:51:24 dedmaus user.warn kernel: ip_conntrack: table full, > dropping packet. > May 4 13:51:24 dedmaus user.warn kernel: ip_conntrack: table full, > dropping packet. > May 4 13:51:24 dedmaus user.warn kernel: ip_conntrack: table full, > dropping packet. > May 4 13:51:25 dedmaus user.warn kernel: ip_conntrack: table full, > dropping packet. > May 4 13:51:29 dedmaus user.warn kernel: NET: 152 messages suppressed. > May 4 13:51:29 dedmaus user.warn kernel: ip_conntrack: table full, > dropping packet. > May 4 13:51:34 dedmaus user.warn kernel: NET: 193 messages suppressed. > May 4 13:51:34 dedmaus user.warn kernel: ip_conntrack: table full, > dropping packet. > May 4 13:51:39 dedmaus user.warn kernel: NET: 227 messages suppressed. > > ...ad infinitum with the number of messages suppressed per 5 sec > increasing until the router crashes. > > > > On Mon, Mar 18, 2013, at 06:18 PM, torsion at ftml.net wrote: >> I'm also seeing occasional messages like this on the Pi (it never lasts >> long): >> >> 18:13:24 [ARM_NOTICE] Relay resumed >> 18:13:18 [ARM_NOTICE] Relay unresponsive (last heartbeat: Mon Mar 18 >> 18:13:04 2013) >> 17:28:43 [ARM_NOTICE] Relay resumed >> 17:28:38 [ARM_NOTICE] Relay unresponsive (last heartbeat: Mon Mar 18 >> 17:28:25 2013) >> 14:12:26 [ARM_NOTICE] Relay resumed >> 14:12:20 [ARM_WARN] Deduplication took too long. Its current >> implementation has difficulty handling large logs so disabling it to >> keep the interface responsive. >> 14:12:20 [ARM_NOTICE] Relay unresponsive (last heartbeat: Mon Mar 18 >> 14:12:06 20 >> >> On Mon, Mar 18, 2013, at 01:00 PM, torsion at ftml.net wrote: >> > Hi there, I just joined the mailing list and apologized if this has >> been >> > discussed before. I did find discussion of a similar issue in January >> > 2013's archive: >> > >> > https://lists.torproject.org/pipermail/tor-relays/2013-January/001809.html >> > >> > It's important to note that I believe I've seen (but didn't save logs) >> a >> > couple "circuit creation burst" events on my established relay (about >> > 5Mbps, stable, guard, non-exit) which was mostly able to handle it >> > without crashing as it has plenty of RAM and the above-mentioned >> > messages - "Your computer is too slow to handle this many circuit >> > creation requests! Please consider using the MaxAdvertisedBandwidth >> > config option or choosing a m ore restricted exit policy." - appear >> only >> > with the relay is under load for other reasons AND a large number of >> > circuits are being suddenly created. >> > >> > I wondered if this was some kind of DOS attempt but didn't think much >> of >> > it because my fast relay continued working fine. >> > >> > However, I've just set up a Raspberry Pi, the 512MB model, as a relay >> on >> > a slower connection. Here are the relevant settings on this relay: >> > >> > RelayBandwidthRate 130 KB >> > RelayBandwidthBurst 340 KB >> > >> > The Pi has a fairly slow CPU, so I'd occasionally get messages about >> log >> > deduplication being too slow or something, but didn't think much of >> it. >> > I finally got the relay up and left it up for over 24 hours. When I >> > woke up this morning it had crashed. Here are the relevant log >> messages >> > - note the huge jump in number of circuits between 22:35 and 04:35 >> > (maybe I got the Stable flag), then the storm of circuit open requests >> > starting at 05:49. Eventually I believe the Pi ran out of memory and >> > killed the tor process. >> > >> > What's very interesting here is that my fast VPS relay with a >> > RelayBandwidthRate over 5x faster is almost never handling much more >> > than 1000 circuits, so why all of a sudden the demand on the Pi that's >> > advertising a lower bandwidth rate? >> > >> > Mar 17 22:35:00.000 [notice] Heartbeat: Tor's uptime is 1 day 0:00 >> > hours, with 26 circuits open. I've sent 974.13 MB and received 969.92 >> > MB. >> > Mar 18 04:35:00.000 [notice] Heartbeat: Tor's uptime is 1 day 6:00 >> > hours, with 972 circuits open. I've sent 1.61 GB and received 1.59 GB. >> > Mar 18 05:49:44.000 [warn] Your computer is too slow to handle this >> many >> > circuit creation requests! Please consider using the >> > MaxAdvertisedBandwidth config option or choosing a more restricted >> exit >> > policy. >> > Mar 18 05:49:44.000 [warn] Failed to hand off onionskin. Closing. >> > Mar 18 05:50:44.000 [warn] Your computer is too slow to handle this >> many >> > circuit creation requests! Please consider using the >> > MaxAdvertisedBandwidth config option or choosing a more restricted >> exit >> > policy. [5817 similar message(s) suppressed in last 60 seconds] >> > Mar 18 05:52:30.000 [warn] Your system clock just jumped 101 seconds >> > forward; assuming established circuits no longer work. >> > Mar 18 05:53:51.000 [warn] Your computer is too slow to handle this >> many >> > circuit creation requests! Please consider using the >> > MaxAdvertisedBandwidth config option or choosing a more restricted >> exit >> > policy. [1055 similar message(s) suppressed in last 60 seconds] >> > Mar 18 05:55:14.000 [warn] Your computer is too slow to handle this >> many >> > circuit creation requests! Please consider using the >> > MaxAdvertisedBandwidth config option or choosing a more restricted >> exit >> > policy. [329 similar message(s) suppressed in last 60 seconds] >> > >> > I'd like to figure out just how much the Raspberry Pi is capable of, >> > because it could be a cheap way to build out the relay network by >> people >> > who want to donate bandwidth - but of course it needs to be stable, >> and >> > something about my setup is not. >> > >> > Also: >> > >> > Mar 16 20:55:33.000 [notice] No AES engine found; using AES_* >> functions. >> > >> > I have no idea if the Broadcom BCM2835 SoC (ARM1176JZF-S CPU) in the >> Pi >> > has any AES capability, but it'd be great to find out. > > >

4 3

Running Obfsproxy on a Raspberry Pi
by Richard Budd 02 Jun '13

02 Jun '13

Don't know how common this is but I've had a Pi running for 35 days 6 hours (so far). With over 80GB transferred on my half assed comcast cable connection. Not bad for $25 a credit card sized board sitting in a cardboard box in my broom closet. I think I'm going to give one to everyone of my family members preloaded with Tor. Plug it into their cable router and let it run. That would be another 4 bridges added to the total. If we could get another 100 people to do this it it might be a good way to add capacity with very little cost or power use. BTW, I have nothing to do with the Pi foundation, I just like playing around with them. Richard

2 5

Re: [tor-relays] Issues with vidalia-obfsproxy-bridge-bundle-0.2.4.12-alpha-0.3.3-alpha on Windows
by Asa Rossoff 01 Jun '13

01 Jun '13

From: Asa Rossoff [mailto:asa@lovetour.info] Sent: Saturday, June 01, 2013 4:33 AM To: 'tor-talk(a)lists.torproject.org' Cc: 'tor-relays(a)lists.torproject.org' Subject: Issues with vidalia-obfsproxy-bridge-bundle-0.2.4.12-alpha-0.3.3-alpha on Windows [A. R.] . > 5) Sometimes, problem #4 never occurs: Tor simply stops generating log messages and I dissappear from the consensus (onionoo status=down). This also takes some hours. This is the state at this moment as I write this message, and looking at the log, there is nothing unusual reported (level=notice/warning) (if beneficial, I can learn how to have Tor generate a full debug log when run directly without Vidalia). Tor is still functioning as a client -- without log messages to the terminal, at least for a longstanding existing connection. [A. R.] Interesting development just now: I went to the Tor console window that has shown no activity for many hours (andonionoo shows me down). I was about to close it and restsart Tor, but on a whim hit the ESC keep instead. Tor immediatiately came back to life, ebginning with the following message "[warn] Your system clock just jumped 64376 seconds forward; assuming established circuits no longer work." and restested it's OR port (which succeeded quickly). I guess it's working again.. will leave it for now..strange trange strange. I guess I can run it from a batch file that pipes a huge file with ESC's into it.. might keep it alive.. Thanks again if any insights; Asa

1 0

Issues with vidalia-obfsproxy-bridge-bundle-0.2.4.12-alpha-0.3.3-alpha on Windows
by Asa Rossoff 01 Jun '13

01 Jun '13

Dear tor-talk and tor-relays experts: Hi, The solution to my problems is probably obvious: "well, downgrade" (or perhaps, if skilled or daring, try compiling the latest code), however: I'd like to report that the current obfsproxy bridge bundle for Windows seems to be essentially unusable on my system. Windows 7 Ultimate x64, first gen i3/2 core/4 thread mobile system. I think this is the first Windows obfsproxy bridge bundle released. Previously I used the 0.2.4.11 pyobfsproxy browser bundle and simply reconfigured it as a bridge, which seemed to work fine and be very stable. My initial difficulties were all my own: configuration file changes between the two versions and filename changes (pyobfsproxy -> obfsproxy). Solved them, to my knowledge. 1) Then, Tor always crashed on startup, which may have been a timing issue as this was mostly/only observed with Vidalia configured to startup with the system/login. This "fixed itself" (more evidence it may have been a timing issue with differing startup/login timing). However, I think this is a bug, as startup with the system should just work (I had no problems with 0.2.4.11). 2) Then, although Tor didn't crash on startup anymore, there were numerous error messages (unfortunately, I don't recall them) in the Tor log when starting with the system, and I think it was not fully working anyway. 3) Maybe a showstopper: in between 20 minutes and and an hour, Vidalia either lost communication with Tor, or Tor was failing in some way. Vidalia did not know it had lost communication, but there were no more log messages, and all the other information tabs either reported nothing, or in the case of the network map -- mapped nothing but showed some connections. Nothing happened if I tried to terminate connections from the interface. Clicking the "Stop Tor" button and then "Start Tor" resolved the issue -- for 20-60 minutes. My solution: stop using Vidalia, and just run Tor directly. I made a hardlink to the torrc file in Vidalia's appdata folder, from Tor's so they could both find it (there is probably a command-line switch or two I could have used :) -- I didn't check!) 4) Now, things were better. Next problem: periodically, usually after some hours of running fine, the Tor log began showing errors re. the cached-microdescs file. First, a permission/file access error, and then hundreds of other errors related to building the micro descriptor cache would be generated every some minutes. Actually---I also observed this problem the first time, even running under Vidalia, that Tor ran without crashing on startup. My solution attempt: delete the file. It was locked. When I closed Tor --- the lock was released [Tor was tied to the file lock] and I could delete the file. I restarted Tor and it worked fine for some hours before repeating this behavior. Once this happens, my node goes down. I think it has to do more bootstrapping work on restart when this file is lost, if I understand correctly. 5) Sometimes, problem #4 never occurs: Tor simply stops generating log messages and I dissappear from the consensus (onionoo status=down). This also takes some hours. This is the state at this moment as I write this message, and looking at the log, there is nothing unusual reported (level=notice/warning) (if beneficial, I can learn how to have Tor generate a full debug log when run directly without Vidalia). Tor is still functioning as a client -- without log messages to the terminal, at least for a longstanding existing connection. 6-sort of) I'll note in passing---in the off-chance it might be an issue on the Tor end, and in the other off-chance that an issue on "my end" is adding to Tor's instability. I suppose in both cases it might be called a Tor bug if in deed relevant at all, as I would expect Tor to handle it gracefully: I have my OR address specified using a DNS name (a dynamic DNS name). I run a dynamic DNS updater that either polls for changes every 5 minutes, or else Tor does a lookup every 5 minutes, because Tor is reporting IP address changes sometimes as frequently as every 5 minutes, between my real IP (which has not changed) and an IP on an entirely different network owned by an unaffiliated ISP!. This gererates frequent ORPort reachablity tests, which are always done, for some reason, on the correct IP address. The incorrect address has a long tracert to another state before it dwindles off to no-reporting private network routers, and a reverse nslookup indicates it is on a cellular address block. --- I will have to study this issue on my own, most likely (bug in updater, my dynamic dns service, etc. - it may be the last known address on one of the configured interfaces on my coputer - which is down -- for when I use my phone's connection on occasion.) I can simply try returning to my old batch file that worked great as a dynamic ip updater, and can the fancy but possibly buggy executable in the system tray. 7) Minor issue: ORPort reachability tests do not always succeed in the 20 minutes alotted (but I can perform a telnet test from a foreign server to the port and get a connection). The tests are repeated and eventually succeed. On the other hand, sometimes the test succeeds within seconds. I have trouble understanding why they would fail (or why they would take more than 30 seconds) unless it is an issue where other nodes refuse to cooperate in reachability testing or make it a low priority. This is probably not a version-specific issue, but might be a design issue. 8) This issue comes chronologically early in the sequence of problems I experienced: Some controls in the Vidalia 0.2.4.12-0.3.3 Settings dialog that I tried didn't save the values, So I relied on manually editing the config files. -------------- If you've read this far, thanks! I will soon--time permitting--and lacking other advice--try the non-obfsproxy bridge bundle (0.2.4.12-alpha-0.2.21) supplemented with the obfsproxy executable from alpha 0.3.3. If that fails, I will go back to the 0.2.4.11 browser bundle configured as a bridge. When I achieve a stable combination, I might try a tool for installing anything as a service and try to run the Tor executable as a service. Seems like Vidalia should be able to talk to it, but I might have to overcome automatic Vidalia behavior like trying to start Tor itself and exiting if the Tor process is killed, etc. ---------------- There are many great minds here, so if anyone has any pointers or links o a page with the info I need, that would be great. Otherwise, consider this a bug report and let me kow if I can aid in debugging. I want to support the Tor network in such ways as I can. Thanks, Asa

1 0

Re: [tor-relays] [tor-talk] Skype banned tor-nodes?
by Mike Perry 29 May '13

29 May '13

Cross-posting to tor-relays to give everyone the heads up - I just added 33033 to the ReducedExitPolicy page for Skype: https://trac.torproject.org/projects/tor/wiki/doc/ReducedExitPolicy Non-relay related discussion should remove tor-relays from the Cc. James Brown: > Hello, people! > > Because many my contacts continue to use Skype and don't want to use any > nice soft (such as jabber, torchat and etc.) I need to use it too. > For protecting of my location and etc. I use it only through Tor (on VM > working under transparently-torifed user). > Some days ago I would be very, very difficalt to connect to skype > through the Tor-net (earlier it was very easy). > As can I see in my Vidalia, my skype application can make resolving > through tor and can connect to hosts through port 443, but conncetions > with port 33033 cannot establish (I can the the trying to open it and > after it it closed immediatly). > As can I see the skype doesn't work without that connections which now > establishes very rarely. > I have 2 versions: > a) the Skype team ban the Tor-connections to skype (but I can still > connect to skype to ports 443 and I can to browse their web-site and > even log in to it through the Tor without any problems) > or > b) many Tor-exits don't let the above-mentioned port (33033) in their > ExitPolicy. > It is not good from the Skype team if they do the first. > If the second hypothesis is right I ask owners of Exit-nodes, if it > possible, to let that port in their ExitPolicies. Not sure if that's actually the problem, but if the only way you can get to Skype is to use a Bittorrent-supporting exit, it certainly seems like a possibility. Thanks for the heads up James! -- Mike Perry

1 0

obfsproxy on freebsd
by Peter Kasper 28 May '13

28 May '13

hi, i'm trying to run obfsproxy on freebsd 9.1, i have installed packages obfsproxy-0.1.4,1 and tor-devel-0.2.4.12.a log file: May 24 16:36:29.000 [notice] Your OpenSSL version seems to be 0.9.8y. We recommend 1.0.0 or later. May 24 16:36:52.000 [notice] To correct this, use a more recent OpenSSL, built without disabling any secure ciphers or features. May 24 16:36:53.000 [warn] The communication stream of managed proxy '/usr/local/bin/obfsproxy' is 'closed'. Most probably the managed proxy stopped running. This might be a bug of the managed proxy, a bug of Tor, or a misconfiguration. Please enable logging on your managed proxy and check the logs for errors. May 24 16:36:53.000 [notice] Failed to terminate process with PID '12021' ('No such process'). when i start obfsproxy directly, there is only obfs2 protocol 2013-05-25 21:09:51 [notice] Starting (obfsproxy version: 0.1.4 (git-94ebc4c3edf1e3e5)). Usage: obfsproxy [obfsproxy_args] protocol_name [protocol_args] protocol_options protocol_name ... * Available protocols: [dummy] [obfs2] any idea? peter ---------- ZPS 14 za 16,90 EUR - vypredaj internetoveho skladu pokracuje! SUPER AKCIA! Registracia nove .EU domeny len za 1,99 EUR! Najdite si strateny alebo ukradnuty telefon programom Zoner AntiVirus Free pre Android Domeny .SK len za 12.29 EUR Ponuka z vydavatelstva Zoner Press Mam rad Android - klikni tu

2 2

Increasing nofiles when running Tor-Relay in a chroot environment
by Sven Reissmann 28 May '13

28 May '13

Hi, I'm running Tor inside a chroot that I built long ago, using this guide: https://trac.torproject.org/projects/tor/wiki/doc/TorInChroot For some weeks my traffic limitations are gone, so I' d like to increase the allowed traffic on my relay. Problem is, that I also need to increase the maximum number of open file descriptors, tor may use. I did this in the limits.conf configuration file, but it has no effect in the chroot environment. Is anyone running a similar setup or has a hint how it could be solved? Regards, Sven. -- PGP Schluessel : https://0x80.io/files/key.asc Anonyme Antwort: https://privacybox.de/srm.msg

1 0

Several IP address on the same box
by Torry Torah 24 May '13

24 May '13

Hi, I have two network interfaces on my server and thus 2 IP addresses. Since tor uses only a very small fraction of my resources (I'm running a bridge), I was wondering if I could put up two bridges, one on each interface. I didn't find anything in the docs about that. Many thanks, -- Torry

2 1

Is TOR using more than just OrPort and DirPort?
by Daniel Wu 13 May '13

13 May '13

I just started a relay (non-exit node, not running a client myself), on Windows, using the latest Vidalia Relay Bundle. Looks like from the configuration, given that I'm running in relay mode, there should be only two ports used by Tor (OrPort and DirPort), right? However, when I run "netstat -ano", I see that the Tor process is using up a lot more ports. There are these connections, from 127.0.0.1 back to itself. Some sort of internal process used by Tor? Not as concerned about these, since these are internal. But still curious. TCP 127.0.0.1:63417 127.0.0.1:63418 ESTABLISHED TCP 127.0.0.1:63418 127.0.0.1:63417 ESTABLISHED TCP 127.0.0.1:63419 127.0.0.1:63420 ESTABLISHED But then I see connections like these: TCP 192.168.1.202:55049 174.136.105.86:9001 ESTABLISHED TCP 192.168.1.202:56804 37.128.208.46:9002 ESTABLISHED TCP 192.168.1.202:56896 171.25.193.9:80 ESTABLISHED TCP 192.168.1.202:57113 109.232.224.74:9001 ESTABLISHED TCP 192.168.1.202:57206 91.227.249.44:9001 ESTABLISHED TCP 192.168.1.202:57221 67.164.46.197:9001 ESTABLISHED TCP 192.168.1.202:57253 128.31.0.34:9101 ESTABLISHED TCP 192.168.1.202:57259 204.124.83.132:587 ESTABLISHED TCP 192.168.1.202:57260 128.232.18.57:9001 ESTABLISHED TCP 192.168.1.202:57309 204.124.83.131:443 ESTABLISHED TCP 192.168.1.202:57331 81.24.98.236:9001 ESTABLISHED These appear to be the actual Tor relay traffic (192.168.1.202 is my computer). Why are these using ports in the 55000+ range, when I specified my OrPort to be a singular value (in my case, 9031)? I would like to know the port ranges used by Tor for relay traffic, so I can use my dd-wrt to set the QoS by specifying these Tor port ranges. Thanks, DW

4 6

"Localized" TOR exit notices?
by Conrad Hoffmann 07 May '13

07 May '13

Hi all, I recently decided I should run a TOR exit node. I am new to the list and still in the process of setting things up. One thing I noticed is that the TOR exit notice [1] contains a US-specific section (it even says so in the comments). Is there a resource anywhere as to how this paragraph might look in other countries? I am located in Germany for example, and I suppose a short mention of the TMG [2], specifically § 8 and § 15 would be suitable replacement for the above mentioned paragraph. I guess that even if not in the git repo, at least a collection in the wiki might be a neat idea? Or did I just not find it? Cheers and thanks for any hints, Conrad [1] https://gitweb.torproject.org/tor.git/blob/HEAD:/contrib/tor-exit-notice.ht… [2] http://www.gesetze-im-internet.de/tmg/index.html (in german)

2 1