Launcher run works different than enter and run command

AI-generated summary

Users are discussing differences in behavior between ./launcher enter app and ./launcher run app "command" when executing commands in a Discourse container. iurisilvio reports that ./launcher run fails with a Redis connection error, while ./launcher enter works as expected. pfaffman confirms ./launcher run works in some cases, suggesting a possible issue with the environment. iurisilvio resolves the issue using docker exec, but expects ./launcher run to be the standard method. EricGT investigates the launcher run command and discovers it uses docker run with specific flags, noting the complexity and potential for issues, especially with environment setup. The discussion highlights that ./launcher run may not fully replicate the context of ./launcher enter, and users seek a more reliable and straightforward method.

Entering app context and running my command works as expected.

/var/discourse# ./launcher enter app
/var/www/discourse# discourse disable_restore
Restore are now forbidden. Enable them with `enable_restore`

I expected the ./launcher run app "discourse disable_restore" to be the same thing, implicitly changing to the docker context, but it fails with an weird redis connection refused.

/var/discourse# ./launcher run app "discourse disable_restore"
Failed to report error: Error connecting to Redis on localhost:6379 (Errno::ECONNREFUSED) 2 Error connecting to Redis on localhost:6379 (Errno::ECONNREFUSED) subscribe failed, reconnecting in 1 second. 
Call stack ["/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis/client.rb:344:in `rescue in establish_connection'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis/client.rb:328:in `establish_connection'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis/client.rb:99:in `block in connect'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis/client.rb:291:in `with_reconnect'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis/client.rb:98:in `connect'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis/client.rb:274:in `with_socket_timeout'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis/client.rb:131:in `call_loop'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis/subscribe.rb:43:in `subscription'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis/subscribe.rb:12:in `subscribe'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis.rb:2824:in `_subscription'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis.rb:2192:in `block in subscribe'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis.rb:45:in `block in synchronize'", "/usr/local/lib/ruby/2.5.0/monitor.rb:226:in `mon_synchronize'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis.rb:45:in `synchronize'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/redis-4.0.1/lib/redis.rb:2191:in `subscribe'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/message_bus-2.1.6/lib/message_bus/backends/redis.rb:337:in `global_subscribe'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/message_bus-2.1.6/lib/message_bus.rb:533:in `global_subscribe_thread'", "/var/www/discourse/vendor/bundle/ruby/2.5.0/gems/message_bus-2.1.6/lib/message_bus.rb:481:in `block in new_subscriber_thread'"] 
bundler: failed to load command: script/discourse (script/discourse)
Redis::CannotConnectError: Error connecting to Redis on localhost:6379 (Errno::ECONNREFUSED)

Is it expected behaviour?

I’m running from master (discourse 42aefc3 and discourse_docker 3377f26).

It worked for me just now.

root@lh:/var/discourse#` ./launcher run web_only "discourse disable_restore"`
Restore are now forbidden. Enable them with `enable_restore`
root@lh:/var/discourse# ./launcher run web_only "discourse enable_restore"
Restore are now permitted. Disable them with `disable_restore`

But maybe there’s something different in a web-only vs single-container solution. You might try

 ./launcher run app  "su discourse discourse disable_restore"

Tried your solution but does not work. :frowning:

/var/discourse# ./launcher run app "su discourse discourse disable_restore" discourse is not in the sudoers file. This incident will be reported.

Hmm. Well, I’ll try on a single container install today if I get a chance.

I solved my issue with another @pfaffman answer: Running commands inside discourse container? - #5 by pfaffman

docker exec -w /var/www/discourse -i app discourse enable_restore

Works for me, but I expected ./launcher run app "command" to be the way to do it.

Thanks!

Glad my answer worked! The ./launcher run command is pretty new, but I’d expect it to work too.

Hi all!

I ran into the same issue. I’d like to run ./launcher to instruct enabling restore. What is the correct way to do it?

So instead of

vagrant ssh -c 'cd /vagrant;sudo ./launcher run discourse discourse enable_restore'

it works for me (as stated above):

vagrant ssh -c 'sudo docker exec -w /var/www/discourse -i discourse discourse enable_restore'

but I expect something more straightforward. I simply don’t wan’t to go under the hood and stay “safe” using the standard commands. I though launcher is for that purpose.

Glad you said that.

I know the post is from 2018, but that was probably not a simple command to get right given the expected users.

I have been spending the last few days taking the launcher script apart piece by piece and am now on the run command. When working on the launcher commands I also try to recreate the same steps using the docker commands used.

For run I knew there were two docker commands that use run

docker container run
docker run

but not being a Docker expert I can not explain to the rubber duck what the differences are or if there are none.

So when I tried a simple ls, it took so long I almost took a beverage break. It did return a correct ls result so that is when I took a look at exactly what launcher run was doing.

When I saw this

  (exec $docker_path run --rm --shm-size=512m $user_args $links "${env[@]}" -e DOCKER_HOST_IP="$docker_ip" -i -a stdin -a stdout -a stderr $volumes $run_image \
    /bin/bash -c "$run_command") || ERR=$?

I could not believe it. I haven’t asked why it is that way because sometimes the answer is in a post if you look for it, but also if I don’t find the answer I will be much better informed.

** EDIT **

Related post: Trying to understand launcher as a series of Docker commands

Result of running launcher run app ls as noted in related post, i.e.

groot@galaxy:/var/discourse$ sudo bash -x launcher run app ls 2> ~/traces/ls_01
eric@swi-prolog:/var/discourse$ grep /usr/bin/docker ~/traces/ls_01

Otuput:

+ docker_path=/usr/bin/docker
+ '[' -z /usr/bin/docker ']'
++ /usr/bin/docker info
+ /usr/bin/docker info
++ /usr/bin/docker --version
++ /usr/bin/docker images
++ /usr/bin/docker run -i --rm -a stdout -a stderr discourse/base:2.0.20200512-1735 echo working
++ /usr/bin/docker info --format '{{.DockerRootDir}}'
+ '[' -z /usr/bin/docker ']'
++ /usr/bin/docker --version
+++ /usr/bin/docker run --rm -i -a stdin -a stdout discourse/base:2.0.20200512-1735 ruby -e 'require '\''yaml'\''; puts YAML.load(STDIN.readlines.join)['\''templates'\'']'
++++ /usr/bin/docker run --rm -i -a stdin -a stdout discourse/base:2.0.20200512-1735 ruby -e 'require '\''yaml'\''; puts YAML.load(STDIN.readlines.join)['\''templates'\'']'
++ /usr/bin/docker run --rm -i -a stdin -a stdout discourse/base:2.0.20200512-1735 ruby -e 'require '\''yaml'\''
++ /usr/bin/docker run --rm -i -a stdin -a stdout discourse/base:2.0.20200512-1735 ruby -e 'require '\''yaml'\''
++ /usr/bin/docker run --rm -i -a stdin -a stdout discourse/base:2.0.20200512-1735 ruby -e 'require '\''yaml'\''
++ /usr/bin/docker run --rm -i -a stdout -a stdin discourse/base:2.0.20200512-1735 ruby -e 'require '\''yaml'\''; puts YAML.load(STDIN.readlines.join)['\''docker_args'\'']'
++ /usr/bin/docker run --rm -i -a stdout -a stdin discourse/base:2.0.20200512-1735 ruby -e 'require '\''yaml'\''; puts YAML.load(STDIN.readlines.join)['\''volumes'\''].map{|v| '\''-v '\'' << v['\''volume'\'']['\''host'\''] << '\'':'\'' << v['\''volume'\'']['\''guest'\''] << '\'' '\''}.join'
++ /usr/bin/docker run --rm -i -a stdout -a stdin discourse/base:2.0.20200512-1735 ruby -e 'require '\''yaml'\''; puts YAML.load(STDIN.readlines.join)['\''links'\''].map{|l| '\''--link '\'' << l['\''link'\'']['\''name'\''] << '\'':'\'' << l['\''link'\'']['\''alias'\''] << '\'' '\''}.join'
++ /usr/bin/docker run --rm -i -a stdin -a stdout discourse/base:2.0.20200512-1735 ruby -e 'require '\''yaml'\''; puts YAML.load(STDIN.readlines.join)['\''run_image'\'']'
+ exec /usr/bin/docker run --rm --shm-size=512m -e LANG=en_US.UTF-8 -e RAILS_ENV=production -e UNICORN_WORKERS=4 -e UNICORN_SIDEKIQS=1 -e RUBY_GLOBAL_METHOD_CACHE_SIZE=131072 -e RUBY_GC_HEAP_GROWTH_MAX_SLOTS=40000 -e RUBY_GC_HEAP_INIT_SLOTS=400000 -e RUBY_GC_HEAP_OLDOBJECT_LIMIT_FACTOR=1.5 -e DISCOURSE_DB_SOCKET=/var/run/postgresql -e DISCOURSE_DB_HOST= -e DISCOURSE_DB_PORT= -e LETSENCRYPT_DIR=/shared/letsencrypt -e DISCOURSE_HOSTNAME=discourse.galaxy.org -e DISCOURSE_DEVELOPER_EMAILS=groot@galaxy.org -e DISCOURSE_SMTP_ADDRESS=smtp.mailgun.org -e DISCOURSE_SMTP_PORT=587 -e DISCOURSE_SMTP_USER_NAME=groot@mg.galaxy.org -e DISCOURSE_SMTP_PASSWORD= I_am_Groot -e LETSENCRYPT_ACCOUNT_EMAIL=groot@galaxy.org -e DOCKER_HOST_IP=NNN.MMM.OOO.PPP -i -a stdin -a stdout -a stderr -v /var/discourse/shared/standalone:/shared -v /var/discourse/shared/standalone/log/var-log:/var/log local_discourse/app /bin/bash -c ls
``