1

I'm getting below error message in error.log of ejabberd when i try to authenticate through an external python script (Tried with java also).

External script is succesfully receiving inputs from ejabberd and processing it how it should. But it takes a long time to receive the input in external script , by that time ejabberd gets timedout.

Error Log:

2023-03-18 18:01:27.127906+00:00 [error] <0.925.0>@ejabberd_auth_external:failure/4:103 External authentication program failed when calling 'check_password' for vipin@xmpp.mydomain.org: timeout   
2023-03-18 17:46:12.699876+00:00 [error] <0.667.0>@supervisor:do_restart/3:736 SUPERVISOR REPORT:
            supervisor: {local,'extauth_pool_xmpp.mydomain.org'}
            errorContext: child_terminated
            reason: normal
            offender: [{pid,<0.668.0>},
                       {id,'extauth_pool_xmpp.mydomain.org_1'},
                       {mfargs,{extauth,start_link,
                                        ['extauth_pool_xmpp.mydomain.org_1',
                                         "/usr/bin/python3 /home/ejabberd/external-auth.py"]}},
                       {restart_type,permanent},
                       {significant,false},
                       {shutdown,5000},
                       {child_type,worker}]

External auth configuration in ejabberd.yml :

auth_method: external
extauth_program: "/usr/bin/python3 /home/ejabberd/external-auth.py"
extauth_pool_size: 1

I'm using ejabberd/ecs docker image. Copied python script inside the custom docker image created from ejabberd/ecs.

Any help much appreciated

Vipin CP
  • 3,642
  • 3
  • 33
  • 55
  • You should mention what exact image version are you using. Also, there's an example python script, you should try with it and mention in your problem description whether that one fails too, or works correctly: https://docs.ejabberd.im/developer/guide/#external – Badlop Mar 20 '23 at 11:23
  • @Badlop I am using `ejabberd/ecs:latest `docker image which has `ejabberd 23.1.0` version. Tried with same python script mentioned in the website also. – Vipin CP Mar 20 '23 at 12:20

1 Answers1

1

I created a container from image ejabberd/ecs:latest

Copied your three configuration lines, and the example script from the documentation, ejabberd complains at start:

main_1  | sh: exec: line 0: /usr/bin/python3: not found
main_1  | 2023-03-20 17:02:10.958586+00:00 [error] Failed to start external authentication program '/usr/bin/python3 /home/ejabberd/external-auth.py'
main_1  | 2023-03-20 17:02:10.958801+00:00 [error] SUPERVISOR REPORT:
main_1  |     supervisor: {local,extauth_pool_localhost}
main_1  |     errorContext: child_terminated
main_1  |     reason: normal
main_1  |     offender: [{pid,<0.660.0>},
main_1  |                {id,extauth_pool_localhost_1},
main_1  |                {mfargs,{extauth,start_link,
main_1  |                                 [extauth_pool_localhost_1,
main_1  |                                  "/usr/bin/python3 /home/ejabberd/external-auth.py"]}},
main_1  |                {restart_type,permanent},
main_1  |                {significant,false},
main_1  |                {shutdown,5000},
main_1  |                {child_type,worker}]
main_1  | 
main_1  | sh: exec: line 0: /usr/bin/python3: not found

So I add python3:

apk add python3

Now ejabberd starts correctly. When I try to login to an existing account, it logins correctly:

main_1  | 2023-03-20 17:05:04.394536+00:00 [info] (<0.720.0>)
          Accepted connection [::ffff:172.18.0.1]:38496
          -> [::ffff:172.18.0.5]:5222
main_1  | 2023-03-20 17:05:04.502243+00:00 [info] (tls|<0.720.0>)
          Accepted c2s PLAIN authentication for admin@localhost
          by external backend from ::ffff:172.18.0.1
main_1  | 2023-03-20 17:05:04.531838+00:00 [info] (tls|<0.720.0>)
          Opened c2s session for admin@localhost/tka1

I use this docker-compose.yml

version: '3.7'

volumes:
  main_conf:
    name: main_conf
  main_logs:
    name: main_logs
  main_data:
    name: main_data
  main_uplo:
    name: main_uplo

services:

  main:
    image: ejabberd/ecs:latest
    environment:
      - ERLANG_NODE_ARG=ejabberd@main
      - ERLANG_COOKIE=dummycookie123
      - CTL_ON_CREATE=register admin localhost asd
      - CTL_ON_START=stats registeredusers ;
                     status
    command: ["foreground"]
    healthcheck:
      test: netstat -nl | grep -q 5222
      start_period: 5s
      interval: 5s
      timeout: 5s
      retries: 120
    ports:
      #- "4369-4399:4369-4399"
      - "5222:5222"
      - "5269:5269"
      - "5280:5280"
      - "5443:5443"
    volumes:
      - main_conf:/home/ejabberd/conf
      - main_data:/home/ejabberd/database
      - main_logs:/home/ejabberd/logs
      - main_uplo:/home/ejabberd/upload
      - ./bin/ejabberdctl:/home/ejabberd/bin/ejabberdctl:ro
      - ./conf/ejabberd.yml:/home/ejabberd/conf/ejabberd.yml:rw
Badlop
  • 3,840
  • 1
  • 8
  • 9
  • I also had a same setup except i am using network_mode host in docker compose. I can check with the default bridge mode once. Thanks for the reply. – Vipin CP Mar 20 '23 at 20:38
  • It worked when I tried with network mode bridge. Guess some firewall / network related issues which I couldn't really figure out why. I'm not sure if I should accept this answer as right answer , gave an upvote anyways :). – Vipin CP Mar 21 '23 at 11:08