Certain system emails are not sending

[jmturner]

Summary

Discourse is not sending certain system emails. Test emails can be successfully sent from the UI as well as the Ruby scripts and the system successfully sends upgrade reminder emails, so the infrastructure seems correct. However, new users do not get login link emails or email confirmation emails sent. I can see nothing in logfiles to indicate why.

Steps to reproduce

1. Attempt to log in to Discourse
2. Put email/username in form.
3. Click "email me a login link."
4. No email gets sent.

What is the expected correct behavior?

Discourse should send an email to the email of the user, with a login link included.

Relevant logs and/or screenshots

Environment

• Image version / tag: 3.1.0
• Host OS: Debian Bullseye (running in an unprivileged LXC container on Proxmox)

Any logs | docker-compose.yml

``` version: '2.4' services: discourse-app: container_name: discourse-app image: tiredofit/discourse:latest labels: - traefik.enable=true - traefik.http.routers.discourse-example-org.rule=Host(`discourse.example.org`) # REDACTED - traefik.http.services.discourse-example-org.loadbalancer.server.port=3000 volumes: - ./logs:/data/logs - ./data/uploads:/data/uploads - ./data/backups:/data/backups environment: - TIMEZONE=America/New_York - CONTAINER_NAME=discourse-app

  - DB_HOST=postgresdb
  - DB_NAME=REDACTED
  - DB_USER=REDACTED
  - DB_PASS=REDACTED

  - REDIS_HOST=discourse-redis
  - SITE_HOSTNAME=discourse.example.org # REDACTED

  - CONTAINER_ENABLE_MESSAGING=TRUE
  - SMTP_PORT=25
  - SMTP_HOST=postfix
  - SMTP_OPENSSL_VERIFY_MODE=none
  - SMTP_ENABLE_START_TLS=true

  - DEVELOPER_EMAILS=REDACTED@REDACTED

restart: always
networks:

- proxy

  - services
extra_hosts:
  - postgresdb:172.17.0.1
  - postfix:172.17.0.1
ports:
  - 2020:2020
  - 8082:80
  - 3001:3000
  - 10150:10050

discourse-redis: image: tiredofit/redis:7 container_name: discourse-redis volumes: - ./redis:/var/lib/redis environment: - TIMEZONE=America/New_York - CONTAINER_NAME=discourse-redis - CONTAINER_ENABLE_MESSAGING=TRUE - SMTP_HOST=postfix - SMTP_PORT=25 - SMTP_OPENSSL_VERIFY_MODE=none - SMTP_ENABLE_START_TLS=true

networks:
  - services
restart: always
extra_hosts:
  - postgresdb:172.17.0.1
  - postfix:172.17.0.1
ports:
  - 2021:2020
  - 6379:6379
  - 10151:10050

discourse-db-backup: image: tiredofit/db-backup container_name: discourse-db-backup volumes: - ./dbbackup:/backup environment: - TIMEZONE=America/New_York - CONTAINER_NAME=discourse-db-backup - DB_HOST=postgresdb - DB_TYPE=postgres - DB_NAME=REDACTED - DB_USER=REDACTED - DB_PASS=REDACTED - DB_DUMP_FREQ=1440 - DB_DUMP_BEGIN=0000 - DB_CLEANUP_TIME=8640 networks: - services restart: always extra_hosts: - postgresdb:172.17.0.1 - postfix:172.17.0.1 ports: - 2022:2020 - 10152:10050

networks: proxy: services: ~

</details>

<!-- Include anything additional -->
I note that I do have to do a "gem install actionmailer" to get the test_email_settings.rb script to run.

I'm sadly not familiar enough with Discourse or your handy container to know where the issue lies.

### Possible fixes
<!-- If you can, provide details to the root cause that might be responsible for the problem. -->
I am guessing a batch job somewhere (sidekiq?) is not running like it should. I don't know if the unprivileged nature of the LXC container I'm running in is the root cause. I am using the suggested 'nesting=1' which allows containers to run inside containers.