The default misskey docker install guide will place misskey code and your instance data in the same directory. Is something goes wrong on a later time, you may accidentally trash your datas (docker volumes mounted for the postgres database, redis cache...).
Separate the misskey code (git clone...) from you container's volumes, something like this should do:
```
/usr/local/misskey/code
/usr/local/misskey/data
```
You will need a custom yaml file for docker-compose, but it's worth it (see sample yaml file).
Manage your misskey instance containers with docker-compose and the right path for code and datas:
This particular setting in your instance configuration yaml in the `.config/default.yml` file will raises an error when federating with some (secured) mastodon instances, it's not trivial to find out, activate it to avoid such errors:
```
signToActivityPubGet: true
```
## Arm64 specific
Deploying your misskey instance on arm64 using docker is pretty straightforward, you will need to use some generic flavor of docker images instead of specific/pined version from the original docker-compose.yml file (see the exemple yaml file in this repository).
---
# Scripts list:
## backup_postgres.sh
A daily backup for your docker postgresql database, run it daily
It will cycle with the day of week (1..7)
Add it to your crontab, when your instances has its lowest activity, ex:
`0 5 * * * /usr/local/bin/backup_postgres.sh`
## cleanup_docker.sh
Prune all orphaned images, and stopped containers
## get_container_id.sh
Get container ID from name (used in other scripts)
Yolk emojis, from [Bad-blood on Devianart](https://www.deviantart.com/bad-blood/art/Y-o-l-k-s-78118255). This zip file is ready to be imported on misskey (files with names and tags).