Re: [tor-talk] Why obfs4 bridges aren't work in Tails?

[Alan Hiew <alanhiew@xxxxxxxxxxxxxxx>]

Thu, 16 Apr 2015 20:25:08 +0300
s7r <s7r@xxxxxxxxxx>:

> 
> On Debian Wheezy 64bit with Tor 0.2.5.10 and obfs4proxy installed from
> deb.torproject.org/torproject.org obfs4proxy main (via apt-get) 3
> obfs4 bridges out of a total of 31 crashed with no error or warn in
> Tor log.
> 
> /var/log/tor was an empty file
> /var/log/tor/log.1 (log rotation) had in the last lines the regular
> heartbeat notices, with info about total circuits, relayed traffic,
> [... ]
> 
> Simply started the Tor daemon again on these 3 servers and they are up
> and running, no problem. Will keep an eye on this.
> 

I've made some log analisys and (may be) found why obfs4 bridges don't
work correctly.

I've used tcpdump created by that command:

sudo tcpdump ip -n -nn -x | tee tcpdump.txt

and Tor log file located at /var/log/tor/log


The possible reason is IP-packet fragmentation. I
suspect that obfs4-protocol require sending big packet (with "no
fragmentation" flag/option) and router of my local network and(or)
other routers on the connection way can not pass these packets. My
routers answer: "unreachable - need to frag (mtu 1456)". And in Tor
log:


[info] smartlist_choose_node_by_bandwidth(): Empty routerlist passed in to old node selection for rule weight as guard
[info] should_delay_dir_fetches(): Delaying dir fetches (no running bridges known)
[info] compute_weighted_bandwidths(): Empty routerlist passed in to consensus weight node selection for rule weight
as guard

many times.

obfs3, obfs2 and bridges without pluggable transport all are working
correctly.

WBR, Alan Hiew
-- 
tor-talk mailing list - tor-talk@xxxxxxxxxxxxxxxxxxxx
To unsubscribe or change other settings go to
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-talk