because right now I was relying on that script :

#!/usr/bin/env bash
docker ps --filter "name=^selenoid$" --format '{{.Ports}}'
if [ $? -eq 0 ]
then
  echo "Selenoid is already running"
  exit 0
else
  open --background -a Docker
  sleep 5
  curl -s https://aerokube.com/cm/bash | bash && ./cm selenoid start --vnc
  if [ $? -eq 0 ]
  then
    echo "Selenoid started successfully" >&2
    exit 0
  else
   echo "Start selenoid failed" >&2
   exit 1
  fi
fi

simply download latest CM binary from releases page here: https://github.com/aerokube/cm/releases

ok but so I have to parse the release page myself 🙁 ?

(if I want to automate the download)

yes, there were a lot of issues with this script, so we decided to remove it

Thanks for the hint, but I wanted the binary 🙂

It seems to be working like that :

#!/bin/bash

if [ $(docker inspect -f '{{.State.Running}}' "selenoid") = "true" ]
then
  echo "Selenoid is already running"
  exit 0
else
  open --background -a Docker
  sleep 5
  arch=$(echo `uname -s` | awk '{print tolower($0)}')
  cm_url=$(curl --silent "https://api.github.com/repos/aerokube/cm/releases/latest" | jq --arg PLATFORM_ARCH "$arch" -r '.assets[] | select(.name | contains($PLATFORM_ARCH)).browser_download_url')
  curl -L -o cm $cm_url
  chmod +x ./cm
 ./cm selenoid start --vnc
  if [ $? -eq 0 ]
  then
    echo "Selenoid started successfully" >&2
    exit 0
  else
   echo "Start selenoid failed" >&2
   exit 1
  fi
fi

all this stuff isn't really straightforward to use in CI 🙁

shouldn't "cm start" handle the case where selenoid or cm was already started and stop it ? maybe by introducing a "start if needed" option ?

in my case, depending on the continuous integration agent OSX / windows I may run with --use-drivers, so before using docker commands to stop the container I must check wether it is up, and so on...

I am writing more code to get a selenium server up than I wrote in the test itself ^^

CM only starts Selenoid in case when it's not started

2020-11-24T14:24:46.9830600Z x Failed to start: failed to create container: Error response from daemon: Conflict. The container name "/selenoid" is already in use by container "54a31108bdca562b88a935549bebebd8dcefa647c7c9feee91dc9fb43e52719a". You have to remove (or rename) that container to be able to reuse that name.

is this case a bug to report then ?

this should be some previously manually started container

in latest releases CM is starting everything in Docker network called selenoid

hm it seems I am mixing different errors from different sources. our CI tool seems to kill cm when he moves to the next build step...

I am using "&" bash operator to run it in background, is there another way to install it as a daemon ?

cm cannot be run as daemon, it is command line utility to start selenoid, when selenoid started cm exits...

so after "./cm selenoid start --vnc --use-drivers" is executed, selenoid should be up ?