XMPP Service Operators - 2020-04-30


  1. jonas’

    Huxx, I suggest to pipe it through `jq` for prettyprinting

  2. Huxx

    Ok

  3. Maranda

    Apr 30 08:55:07 s2sin558964761310 debug Incoming s2s connection Apr 30 08:55:07 s2sin558964761310 debug Incoming s2s received <stream:stream from='api.observe.jabber.network' xmlns='http://etherx.jabber.org/streams' version='1.0' xml:lang='en' to='lightwitch.org'> Apr 30 08:55:07 mod_s2s debug sending: <?xml version='1.0'?> Apr 30 08:55:07 mod_s2s debug sending: <stream:stream from='api.observe.jabber.network' xmlns='jabber:server' id='cae86f11-6eb3-46dc-9056-7a4553a004c7' xmlns:db='jabber:server:dialback' version='1.0' to='lightwitch.org' xmlns:stream='http://etherx.jabber.org/streams'> Apr 30 08:55:07 mod_s2s debug Sending stream features: <stream:features><starttls xmlns='urn:ietf:params:xml:ns:xmpp-tls'><required/></starttls><bidi xmlns='urn:xmpp:features:bidi'/><compression xmlns='http://jabber.org/features/compress'><method>zlib</method></compression><dialback xmlns='urn:xmpp:features:dialback'><errors/></dialback><sm xmlns='urn:xmpp:sm:2'/><sm xmlns='urn:xmpp:sm:3'/></stream:features> Apr 30 08:55:07 mod_s2s debug sending: <stream:features> Apr 30 08:55:07 s2sin558964761310 debug Received[s2sin_unauthed]: <starttls xmlns='urn:ietf:params:xml:ns:xmpp-tls'/> Apr 30 08:55:07 mod_s2s debug sending: <proceed xmlns='urn:ietf:params:xml:ns:xmpp-tls'> Apr 30 08:55:07 socket debug try to start ssl at client id: 558964761310 Apr 30 08:55:07 socket debug ssl session delayed until writebuffer is empty... Apr 30 08:55:07 s2sin558964761310 debug TLS negotiation started for s2sin_unauthed... Apr 30 08:55:07 socket debug starting ssl handshake after writing Apr 30 08:55:07 socket debug starting handshake... Apr 30 08:55:07 socket debug ssl handshake of client with id:table: 0x558964761310, attempt:1 Apr 30 08:55:07 socket debug ssl handshake of client with id:table: 0x558964761310, attempt:2 Apr 30 08:55:07 socket debug ssl handshake of client with id:table: 0x558964761310, attempt:3 Apr 30 08:55:07 socket debug ssl handshake of client with id:table: 0x558964761310, attempt:4 Apr 30 08:55:07 socket debug ssl handshake of client with id:table: 0x558964761310, attempt:5 Apr 30 08:55:07 socket debug ssl handshake done Apr 30 08:55:07 s2sin558964761310 warn Received invalid XML: </stream:stream> Apr 30 08:55:07 s2sin558964761310 warn Problem was: not well-formed (invalid token) Apr 30 08:55:07 mod_s2s debug sending: <?xml version='1.0'?> Apr 30 08:55:07 mod_s2s debug sending: <stream:stream version='1.0' xmlns='jabber:server' id='' xmlns:stream='http://etherx.jabber.org/streams'> Apr 30 08:55:07 s2sin558964761310 debug Disconnecting (unknown host)[s2sin_unauthed], <stream:error> is: not-well-formed Apr 30 08:55:07 mod_s2s debug sending: <stream:error> Apr 30 08:55:07 mod_s2s debug sending: </stream:stream> Apr 30 08:55:07 s2sin558964761310 info incoming s2s stream api.observe.jabber.network->lightwitch.org closed: not-well-formed Apr 30 08:55:07 s2sin558964761310 debug Destroying incoming session api.observe.jabber.network->lightwitch.org: not-well-formed Apr 30 08:55:07 socket debug try to close client connection with id: 558964761310 Apr 30 08:55:07 socket debug closing delayed until writebuffer is empty Apr 30 08:55:07 socket debug closing client after writing Apr 30 08:55:07 socket debug closing client with id: 558964761310 client to close

  4. Maranda

    🤔 🤔 🤔 🤔 🤔 🤔 🤔

  5. Ge0rG

    CC jonas’

  6. jonas’

    yeah, some odd stuff happens after TLS negotiation

  7. jonas’

    ISTM that it closes the stream before sending the stream header

  8. jonas’

    Maranda, I’d love to debug this further, but https://github.com/mellium/xmpp/issues/31 blocks me from seeing what’s going on

  9. Maranda

    :(

  10. Maranda

    Also on checking s2s shouldn't it at least reopen the stream after starttls? (or it won't see if the service supports SASL External or not)

  11. jonas’

    Maranda, it should, and if it doesn’t, I need to figure out why

  12. jonas’

    but I can’t right now because of #31

  13. jonas’

    it’s a PITA to debug this with binary garbage on my console ;-)

  14. Maranda

    currently it doesn't for sure at least :D

  15. jonas’

    you sure? because I’ve seen external offered to me already

  16. jonas’

    on a non-dialback domain

  17. Maranda

    {"auth_info":{"dialback_offered":true,"sasl_mechanisms":[]},"certificate_expiration":"2020-06-07T18:34:36Z","durations":{"connect":0.643939568,"starttls":0.198006831},"request_id":"QFTlPx_iMNAoplmF@2600:3c03::f03c:92ff:fe24:b7ee","success":true,"total_duration":1.3246845304965973} -> as long as it doesn't present a certificate or it's invalid during the handshake

  18. Maranda

    but I don't see it reopening the stream after STARTTLS on lightwitch.org

  19. jonas’

    that’s odd

  20. jonas’

    because if you try mdosch.de for example, it definitely sees EXTERNAL

  21. jonas’

    (but no dialback)

  22. jonas’

    so far I assumed this was a bug with the SASL/Dialback detection (perfectly possible, the feature negotiation in that library is ... weird to say the least)

  23. Maranda

    Pity, so if a domain supports both it'll fail detection I take (after all to check dialback support it does/should check the stream header, not features)

  24. Maranda

    Pity, so if a domain supports both it'll fail detection I take, after all to check dialback support it does/should check the stream header, not features

  25. jonas’

    huh?

  26. jonas’

    I guess it’s also a variant of https://github.com/mellium/xmpp/issues/47

  27. Maranda

    jonas’, well mdosch.de doesn't look to have mod_dialback loaded.

  28. Maranda

    (hence my previous message)

  29. Maranda

    @ping jabber.fr

  30. Echo1

    Maranda: Pong from jabber.fr in 0.086 seconds

  31. Martin

    > jonas’, well mdosch.de doesn't look to have mod_dialback loaded. Yeah, no dialback there.

  32. Ellenor Malik dialbacks Martin

  33. Martin doesn't pick up