Changes between Initial Version and Version 1 of Ticket #29927


Ignore:
Timestamp:
Mar 30, 2019, 11:02:19 PM (11 months ago)
Author:
mikeperry
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #29927 – Description

    initial v1  
    1 While testing vanguards, I've got some mystery cases client side where circuits are getting closed with END_CIRC_REASON_TORPROTOCOL, but Tor is not emmitting any log lines that correspond to this. This is happening for circuits with purposes CIRCUIT_PURPOSE_C_REND_READY_INTRO_ACKED and CIRCUIT_PURPOSE_C_GENERAL.
     1While testing vanguards, I've got some mystery cases client side where circuits are getting closed with END_CIRC_REASON_TORPROTOCOL, but Tor is not emmitting any log lines that correspond to this, even at debug level.
    22
    3 Additionally, CIRCUIT_PURPOSE_C_INTRODUCE_ACKED circuits are getting closed with a END_CIRC_REASON_FINISHED after receiving an invalid cell, seemingly after they are done being used.
     3This is happening for circuits with purpose CIRCUIT_PURPOSE_C_REND_READY_INTRO_ACKED. Additionally, all circuits seem able to fail during construction with END_CIRC_REASON_TORPROTOCOL, with no Tor log messages even at debug loglevel. Possibly more ntor handshake failures, similar to #29700?
     4
     5Finally, CIRCUIT_PURPOSE_C_INTRODUCE_ACKED circuits are getting closed with a END_CIRC_REASON_FINISHED after receiving an invalid cell, seemingly after they are done being used.
    46
    57See also https://github.com/mikeperry-tor/vanguards/issues/37
     8
     9The vanguards addon now outputs this bug number at INFO log level when this happens.