# Mastodon

> Your self-hosted, globally interconnected microblogging community

<!-- metadata -->
* **Category**: Apps
* **Status**: 1
* **Image**: [`tootsuite/mastodon`](https://hub.docker.com/r/tootsuite/mastodon)
* **Healthcheck**: No
* **Backups**:  No
* **Email**: Yes
* **Tests**: No
* **SSO**: Yes
<!-- endmetadata -->

## Quick start

Mastodon expects secrets to be formatted in a very specific way, so please
choose "No" when prompted to generate secrets for `abra app new mastodon`. The
secrets must be generated outside of `abra` and that is achieved in step 2. See
the [`abra.sh`](./abra.sh) for more.

1. `abra app new mastodon`
1. `abra app cmd <domain> secrets --local`
1. `abra app secret insert <domain> smtp_password v1 <password>`
1. `abra app config <domain>` (uncomment SMTP details)
1. `abra app deploy <domain>`
1. `abra app cmd <domain> streaming setup`

Then, on your host (outside of the containers), you'll need to fix permissions
for the volume (see [#10](https://git.coopcloud.tech/coop-cloud/mastodon/issues/10)):

```
chown -R 991:991 /var/lib/docker/volumes/<domain>_app/_data
```

And finally, within the `app` container, create an admin account:

```
abra app cmd <domain> admin -- <username> <email>
```

## Tips & tricks

### Auto-complete is not working?

Check the sidekiq logs (`/sidekiq/retries`), is a bunch of stuff failing? What
is the error?

If it looks anything like `blocked by: [FORBIDDEN/12/index read-only / allow
delete (api)];` then it might mean that your elastic search service has put
itself into "read-only" state. This could be due to running close to no free
disk space one time. ES doesn't undo this state, even when you have more free
disk space once more, so you need to handle this manually:

```
abra app run <domain> es bash
curl -XPUT -H "Content-Type: application/json" http://localhost:9200/_all/_settings -d '{"index.blocks.read_only_allow_delete": null}'
```

Then head back to the sidekiq retries panel and retry one job. You should see
the ticket of retries go down by one if if passed. Then you can "retry all" and
they should get scheduled & run.