Option to change configuration from /dev/tty to stdout

[chenhayat]

Hi, I am trying to run crossbar status via popen process but it seems that because it's written to /dev/tty it cannot redirect the output and therefore I cannot read the status in python. Is there an option to change the log to write to stdout ?

[oberstet]

yeah, that might be the case, I seem to remember we've added special behavior .. in the context of multi-process (like Crossbar.io) Docker containers stuff gets even more interesting;)

anyways, in general it would be good if you'd posted more: the complete output of crossbar status when run at the interactive terminal of the Crossbar.io you do run when deploying (!) like e.g. from within a Docker container. for a start, this will allow the precise versions of code you actually run - unless you've modified that code and run from there. stuff like that.

[chenhayat]

Thanks for your reply :)
Crossbar version is v21.3.1.
Docker is no used in this case, it's a test that runs crossbar on Linux (20.04) host by executing a Popen python process: crossbar start --config coordinator-cfg.yam and then after that I expected to call crossbar status to make sure it's up.
There is a workaround for that by executing the command in a separate process (multiprocessing.Process), or by executing the crossbar start with multiprocessing.Process and reading the output, but it would be great if there was a direct way to do that :)
It can also be tested with Linux command crossbar status | cat so in this case nothing is printed to console using the pipe.

[oberstet]

Thanks for describing your setup/scenario - that is useful to get a picture. However, it would be good if you could just (also) post the full & unmodified output of crossbar version verbatim here.

Reason: the command was designed and tuned over time to give complete, thorough version info on all relevant software pieces involved when running Crossbar.io. And many of them might be relevant in the context of this kinda pipes and OS and container and whatnot redirection handling ..

[chenhayat]

Sure. that's the output of crossbar version:

  :::::::::::::::::
         :::::          _____                      __
   :::::   :   :::::   / ___/____ ___   ___  ___  / /  ___ _ ____
   :::::::   :::::::  / /__ / __// _ \ (_-< (_-< / _ \/ _ `// __/
   :::::   :   :::::  \___//_/   \___//___//___//_.__/\_,_//_/
         :::::
   :::::::::::::::::   Crossbar v21.3.1

   Copyright (c) 2013-2024 Crossbar.io Technologies GmbH, licensed under AGPL 3.0.

Crossbar.io        : 21.3.1
  txaio            : 23.1.1
  Autobahn         : 21.3.1
    UTF8 Validator : autobahn
    XOR Masker     : autobahn
    JSON Codec     : stdlib
    MsgPack Codec  : msgpack-1.0.8
    CBOR Codec     : cbor-1.0.0
    UBJSON Codec   : ubjson-0.16.1
    FlatBuffers    : flatbuffers-latest
  Twisted          : 24.3.0-EPollReactor
  LMDB             : 1.5.1/lmdb-0.9.31
  Python           : 3.11.9/CPython
  PIP              : 24.2
Frozen executable  : no
Operating system   : Linux-5.15.0-92-generic-x86_64-with-glibc2.31
Host machine       : x86_64
Release key        : RWT906S6dtgOY3NZrHd3uFtXpAPFUc6+bdpADYJRJi4daUaCT5U/0OrO

[oberstet]

I should probably say: we do have a specific option in Crossbar.io to activate use of /dev/tty .. it is pretty deep down buried .. and I also do not understand why it seems to hook in here .. hence I am nagging you for the full version to have a chance to track down whats going on ... anyways, here are some pointers to internal code

crossbar/crossbar/shell/repl.py

Line 164 in 0089c1e

async def repl(old_ctx,

https://github.com/crossbario/crossbar/blob/master/crossbar/_util.py#L47C5-L47C14

---

fwiw, here is my version info on my local box

oberstet@intel-nuci7:~$ cd scm/typedefint/rfminer
oberstet@intel-nuci7:~/scm/typedefint/rfminer$ source .venv/bin/activate
(.venv) oberstet@intel-nuci7:~/scm/typedefint/rfminer$ crossbar version


    :::::::::::::::::
          :::::          _____                 __              _
    :::::   :   :::::   / ___/______  ___ ___ / /  ___ _____  (_)__
    :::::::   :::::::  / /__/ __/ _ \(_-<(_-</ _ \/ _ `/ __/ / / _ \
    :::::   :   :::::  \___/_/  \___/___/___/_.__/\_,_/_/ (_)_/\___/
          :::::
    :::::::::::::::::   Crossbar.io v22.6.1 [19000101-0000000]

    Copyright (c) 2013-2024 Crossbar.io Technologies GmbH. Licensed under EUPLv1.2.

 Crossbar.io        : 22.6.1
   txaio            : 22.2.1
   Autobahn         : 22.7.1
     UTF8 Validator : wsaccel-0.6.2
     XOR Masker     : wsaccel-0.6.2
     JSON Codec     : stdlib
     MsgPack Codec  : msgpack-1.0.4
     CBOR Codec     : cbor2-5.4.3
     UBJSON Codec   : ubjson-0.16.1
     FlatBuffers    : flatbuffers-latest
   Twisted          : 22.8.0-EPollReactor
   LMDB             : 1.3.0/lmdb-0.9.29
   Python           : 3.9.15/CPython
   PIP              : 22.0.4
   NumPy            : 1.23.4
   zLMDB            : 22.6.1
   CFXDB            : 22.6.1
   XBR              : 21.2.1
 Frozen executable  : no
 Operating system   : Linux-5.15.0-122-generic-x86_64-with-glibc2.35
 Host machine       : x86_64
 Release key        : RWRgokUo5rovJw07csyhF67Tw2NOeE7ygo/lxLYK0iF7LVZFxb4O6ti5

(.venv) oberstet@intel-nuci7:~/scm/typedefint/rfminer$ 

[oberstet]

Sure. that's the output of crossbar version:

thanks for that! version info: nothing dubious, looks all reasonable.

IOW: no luck wrt to the issue here:(

anyways, sorry, I don't have time digging deep which would be required, or even adding an option to control it which of course would be possible, busy with other stuff ATM.

[chenhayat]

Ok Thank you for looking into this!