logs archiveIRC Archive / Oftc / #tor / 2009 / December / 13 / 1
fossiiil
tcptraceroute doesn't work with tor :-)
weasel
of course not.
fossiiil
why not?
because the exit gets RST packets, not us?
weasel
because Tor only transports the content within TCP, i.e. the data stream.
it does not transport your tcp segments, even less the ip packets wich would have a TTL set by your traceroute
fossiiil
ok
atagar
gouki, Sebastian: The usr1 signal provides a lot of interesting stats, but not total bandwidth usage. Thus far the only ways I know of doing that is to either sum up BW events (like arm does) or set AccountingMax (in which case tor caps your total usage in order to cap it for the given period)
arma
atagar: no, it does provide total bandwidth usage
it's tucked in the middle of the output
atagar: also, it would be easy to make a getinfo to make tor tell you its internal stat counter
atagar
arma: I'm spotting the average, but not total... still looking
ahhh, when it says "Average bandwidth: 843861/215 = 3924 bytes/sec reading" that's saying the total is 843861, right?
and 215 sec is the uptime
fossiiil
i think that you can also setup netfilter (on Linux) to account data...
         

arma
atagar: yep.
atagar
arma: does it reset at some point like accounting stats and does it persist through a sighup?
arma
i believe it never resets.
(until you restart tor)
Dec 12 18:41:04.024 [notice] Average bandwidth: 497253920088/2169802 = 229170 bytes/sec writing
that's why we made it a uint64
atagar
heh, good call - btw, currently looking at making a proposal to address all that control port wishlist stuff that's accumulated
arma
spiffy. what can i help with?
atagar
let me mull on it for an hour or so (and looking over a rough draft would certainly be appreciated!) - there's always been some uncertainty about how the identifier mappings will work but might be obvious given a tad more coffee ;)
arma
which identifiers?
atagar
mapping between circuit ids, connection ids, ip/ports (see "associate connections to circuits" in the arm TODO)
BarkerJr
does the tor archive site get must traffic?
atagar
but again, give me a little more time - only been looking at it for a few seconds
arma
barkerjr: i don't think we keep track. even if only a few people mirror it, that's many gigs
atagar: ah ha. ok. yeah, i can imagine that one will be tricky to do what all sides want.
BarkerJr
then I won't mirror it :)
fossiiil
Hello Mr. Murdoch! ;-)
arma
barkerjr: heh. if you want to mirror it, i say go ahead. you never know when the current site will get hit by a comet
keb
or a power outage
phobos
archive seems to get very little traffic in general
it will get bursts when someone mirrors it
or downloads everything
keb
or makes an easy to use web script that extracts the info they want
phobos
i should put a README file on there somewhere
to explain how to use rsync to mirror it
and what's in the files
BarkerJr
or why anyone would ever want the files on it? :)
keb
we get someone about once a week in here who needs to know whether a certain ip address was a tor server at a certain date
         

phobos
we have exonerator for that
arma
we need a cgi interface for it
phobos
but neat analytics interface would be helpful
arma
perhaps somebody here wants to work on that? :)
phobos
"data mine tor!"
keb
or maybe once every other week
no time now, but in january if it isnt done i could do something in php
phobos
i'd host it somewhere else, not hosting anything dynamic on archive
;)
but if you write it, great
i'll find a home for it
or use some existing analysis interface
keb
whatis exonerator
phobos
java/python script to pull out IP by time
to answer the question "was this IP a relay on date x,y"
keb
one of these years i will learn python, it seems becoming ubiquitous
nsa
or: arma committed revision 21178 (/website/trunk/en): provide a link for the mailing list reference
keb
in the SIGUSR1 stats, what are App-ward and Exit-ward circuits?
or maybe i should ask, is there doc on the stats
arma
app-ward means towards the client. exit-ward means the other direction.
keb
almost all of them say "0/18437 bytes used on OpenSSL read buffer; 0/18698 bytes used on write buffer."
so i'd guess these stats are for detecting some sort of error condition
arma
you're wasting lots of bytes inside openssl on buffers
nickm patched openssl to get rid of those wasted buffers
keb
ah
arma
- If the user has applied the experimental SSL_MODE_RELEASE_BUFFERS
patch to their OpenSSL, turn it on to save memory on servers. This
patch will (with any luck) get included in a mainline distribution
before too long.
from 0.2.1.1-alpha
keb
i am using a stock debian server, with 0.2.2.6-alpha compiled from source
arma
i suspect you need to patch and rebuild your openssl to get this feature.
keb
i'm not seeing any problems resource-wise, so i can wait for the next package
nsa
or: coderman committed revision 21179 (/torvm/trunk): Minor cleanup of HACKING
polipo: chrisd@seul.org committed patch by Jacob Appelbaum <jacob@appelbaum.net> at Fri, 11 Dec 2009 19:16:34 -0500 (EST) to polipo-commits: [polipo/master] Add GCC-specific hardening flags to our Makefile.
polipo: chrisd@seul.org committed patch by Jacob Appelbaum <jacob@appelbaum.net> at Fri, 11 Dec 2009 19:16:34 -0500 (EST) to polipo-commits: [polipo/master] Add an extra sanity check to avoid memmove segfault
polipo: chrisd@seul.org committed patch by Christopher Davis <chrisd@mangrin.org> at Thu, 10 Dec 2009 00:22:03 -0500 (EST) to polipo-commits: [polipo/master] Fix an integer overflow in processing client connections.
polipo: chrisd@seul.org committed patch by Jacob Appelbaum <jacob@appelbaum.net> at Fri, 11 Dec 2009 19:16:34 -0500 (EST) to polipo-commits: [polipo/master] Add an extra sanity check to avoid memmove segfault
polipo: chrisd@seul.org committed patch by Jacob Appelbaum <jacob@appelbaum.net> at Fri, 11 Dec 2009 19:16:34 -0500 (EST) to polipo-commits: [polipo/master] Add GCC-specific hardening flags to our Makefile.
polipo: chrisd@seul.org committed patch by Christopher Davis <chrisd@mangrin.org> at Thu, 10 Dec 2009 00:22:03 -0500 (EST) to polipo-commits: [polipo/master] Fix an integer overflow in processing client connections.
phobos
woo
keb
yay
arma
weasel: thanks!
weasel
np
phobos
hmm, so I git pull from the polipo repo and get no updates
arma
phobos: these commits happened a few days ago. i just got nsa hooked up with them now.
phobos
ahh
ok then
arma
phobos: http://archives.seul.org/polipo/commits/
currently it's me, chris, jake, nsa, and archiver.
tomorrow i hope to have 100000 followers including celebrities
keb
are yo opening a twitter
murb
git committwit..
nsa
or: coderman committed revision 21180 (/torvm/trunk/build/win32): Update polipo to latest git upstream.
sysop__
hi
sysop2
hi
Sebastian
hi there
sysop2
how do I choose an exit server from a telnet session to the control port (9051?)
if you dont mind me getting to the point.
Sebastian
not at all
fossiiil
hi
i'm always connected via US exit node these days...
sysop2
why for speed?
Sebastian
if you want to choose the path yourself entirely, you need to build a circuit manually. If you only want to control where the exit is, you can use setconf to change the exitnodes config entry and have Tor work with that
sysop2
I used tork.
and its awesome, but doesnt run reliably on my system.
fossiiil
nor on my system
it was crashing
sysop2
yep
Sebastian
In the first case, you'd need to read the control spec or wait for someone who knows how to manually build circuits
sysop2
or using tons of memory
Sebastian
if you only want the other behaviour, you can actually get away with putting the exit nodes in your configuration file, and using Tor that way.
sysop2
I use a thumb drive for swap and the activity light was solid.
I want to do it without restarting tor.
xtoaster
what you 2 are talking about ?
fossiiil
Sebastian: can't you just get list from directory server, choose three node names and tell Controller to use these? :-)
xtoaster
its not like a QA :-D
sysop2
isnt it though?
Sebastian
sysop2: no need to restart Tor, change the config and hup it
changing the config via control port works the same way
sysop2
ok can you give me an example. how do I request the list of exit nodes so I can choose one?
Sebastian
no need, you can parse the cached-consensus file
xtoaster
you cant do it by hand. since the relay id are in base encoded
and you have to use hex encoded relay id in configuration
sysop2
oh. do I have to scrap the list from a webpage somewhere then set my exit to the one I want?
Sebastian
you can
or you can parse the cached-consensus file
sysop2
ok thats sounds easier. now I have the ip address of the exit server I want. how do I change my exit to that ip from a telnet session?
scrapping sounds easier I mean.
dr|z3d
sysop2: mapadress for a single destination. or see strictexitnodes and exitnodes directives in the man.
xtoaster
oh yeah you can convert the relay id to hex code with python. like eddman told me last time. but i forget how to do it :p
sysop2
oh ok I will look at those dr|z3d
thanks
fossiiil
looking at cached-consensus....
xtoaster
I have a question for you sysop2, why not use vidalia?
fossiiil
is it's format documented somewhere?
i see lines beginning r, s, v, w, p
sysop2
btw it sucks that I cant connect to the offical tor irc channel via tor. I know you cant control it. just stating how I feel.
there is vidalia for kubuntu?
fossiiil
:-)
xtoaster
dont know. you can ask in #vidalia
sysop2
ok brb
fossiiil
check official site, i don't know if it's in ubuntu repository, but they surely have .deb package you can install
Sebastian
sysop2: erm, why can't you?
This channel should be open to Tor users
xtoaster
Sebastian: what's vidalia's offical home page?
fossiiil
ok, bye bye everybody, later!
Sebastian
xtoaster: https://www.torproject.org/vidalia/
sysop2
the server rejected me. said I was " Banned autokilled: This host is infected with a keylogger. - Contact support@oftc.net for help. (2009-11-30 20:17:02)"
like I am running windows
Sebastian
sysop2: some Tor exits may have a virus. Choose a different one
dr|z3d
sysop2: The host is likely the exit node.
sysop2
ohhhhh.
ok brb
dr|z3d
"New Identity" from Vidalia will choose a new one.
xtoaster
hey sysop2 its there there is a source tarball on vidalia front page.
i guess that's what you want.
torman2
ok thats better
dont have vidalia for linux installed yet. didnt know it existed. I just used signal NEWNYM
xtoaster
I think there is sth similar to vidalia on kde
i dont know its name
phobos
tork
torman2
tork it not there yet. ttyl
phobos
http://www.anonymityanywhere.com/tork/
Sebastian
phobos: I think torman2 means that tork is crashing for him
torman2
back
linuxguy
hi
ne1 here?
I just install vidalia
ed
it looks just like the windows version, but I dont see how to choose my exit point?
so can someone show me how to choose an exit point via telnet session?
xtoaster
hi you can do as what seb has told you before.
linuxguy
via the config file?
xtoaster
right
and also possiblely with the control port
linuxguy
ok, but it seems kinda kludgey to do it that way when you have a control port.
xtoaster
the point is you can find other relay's id easily
linuxguy
oh I must have missed the part about doing it on the control port.
xtoaster
no need to query from control port or mining data in the cache file.
linuxguy
ok xtoaster maybe the point is I need to learn more about the process. I just wanted quick way to set an exit point so I could script it.
xtoaster
well the most simple way is to use the config file.
linuxguy
I like command line stuff better for network stuff. at least for some things.
how do I hup it? I have heard of that before but dont know how to do it.
xtoaster
ah edmanm showed up
you can talk to him about that.
he is the dev of vidalia :-D
linuxguy
sweet thanks.
edmanm? I dont see that nick?
edeca?
oh in the vid channel ok.
nsa
polipo: chrisd@seul.org committed patch by Fabian Keil <fk@fabiankeil.de> at Sun, 13 Dec 2009 01:39:35 -0500 (EST) to polipo-commits: [polipo/master] Remove dead store in indexDiskObjects().
polipo: chrisd@seul.org committed patch by Fabian Keil <fk@fabiankeil.de> at Sun, 13 Dec 2009 01:39:35 -0500 (EST) to polipo-commits: [polipo/master] In pstrerror(), add missing break for case ESOCKS_REJECT_IDENTD.
polipo: chrisd@seul.org committed patch by Christopher Davis <chrisd@mangrin.org> at Sun, 13 Dec 2009 01:39:35 -0500 (EST) to polipo-commits: [polipo/master] Update CHANGES; fix a bogus entry, and mention Jake's work.
polipo: chrisd@seul.org committed patch by Fabian Keil <fk@fabiankeil.de> at Sun, 13 Dec 2009 01:39:35 -0500 (EST) to polipo-commits: [polipo/master] Remove dead store in parseNetAddress().
polipo: chrisd@seul.org committed patch by Fabian Keil <fk@fabiankeil.de> at Sun, 13 Dec 2009 01:39:35 -0500 (EST) to polipo-commits: [polipo/master] Remove dead store in makeDiskEntry().
polipo: chrisd@seul.org committed patch by Fabian Keil <fk@fabiankeil.de> at Sun, 13 Dec 2009 01:39:35 -0500 (EST) to polipo-commits: [polipo/master] Remove dead store in cachePrintSeparator().
« prev 1 2 3 next »