• how do you use ansible? Is there a good source for roles or playbooks to set up services? I feel like ansible is 30% more headache right now during config.

I write my own playbooks and roles, but often I can just copy paste an existing setup and use it for a new service. For example containers, you can probably write one role once, copy it and modify some variables to set up another container service.
For stuff where there are well maintained community roles (e.g. community.zabbix) just use those and configure with variables.

• how do you deal with motivation loss?

I just don’t work on a part I don’t want to do atm. It’s supposed to mostly be a hobby and as long as my services I care about are running it’s fine.

• how do you deal with the overwhelming amount of choices and information and disciplines (networking, storage, VMS, Linux…) that comes with selfhosting?

I’m on my 2.5th setup now, just choose something and see if it works. If not, see how much it bothers you and what parts you want to migrate.
I’m a big fan of VMs, so I’m using XCP-ng. IMO this makes testing and backups very easy, I just take a snapshot and figure stuff out, no big deal if it breaks.

• how do you find the sweetspot between ease of use, ease of set up, security, redundancy? I feel like I am maybe too pranaoid to loose my data again (dropped a hard drive many years back, I lost all of my projects)

You’re better than 95% of people just by thinking about this. For backups, identify which data you want to back up and do that. If you don’t want to deal with Ansible right now, just set something up manually and automate it later (paste your commands into a readme for reference)
For me, I make sure to backup my Nextcloud data. That included personal photos, files and other hard to replace stuff. Other than that I have daily VM backups to a Hetzner storage box and my NAS. I don’t backup my media on Jellyfin, that’s just not as important.
VMs also make it easy to replace your host. Just install the hypervisor on a new server and restore VMs to it.

• maybe overall, how do you manage your perfectionism?

I guess I’m not a perfectionist. It took me multiple months and monetary incentive (avoid renting two servers) to migrate from my Debian single host setup to VMs years ago.
Some of my Ansible playbooks are “version 1”, where I didn’t know what I was doing. I’m on version 3 now. They still work, I even use some of them occasionally, just haven’t taken the time to migrate them yet.
Maybe you can take a similar approach with some of your services that aren’t that essential and spread out the work more so you can enjoy it when you want to.

I think Memories schedules its scan with the normal cron job for Nextcloud

I use FolderSync to push my photos to Nextcloud. Much more reliable than their app

https://news.ycombinator.com/item?id=8863

Funny you should mention that, Nextcloud originally forked off OwnCloud due to drama.

I think the server is working, it’s basically OwnCloud Infinite Scale but rebranded. Not sure about the client apps, those might be work in progress still. I haven’t really kept up with it either though

Many of the people who worked at OwnCloud Infinite Scale are now at OpenCloud due to disagreements between them and the company which purchased OwnCloud.

You can use host_vars to set different variables per host. You’d still run the same playbook against both hosts, but each has different services activated.

1. Make the folder host_vars in the repo root
2. Make subfolders for every host with their hostname
3. Enable services you want by writing the variables into a yaml file for your host (any file name as long as it ends with .yml
4. Write an Ansible inventory for your hosts
5. Run the playbook with your inventory

Slightly fancier would be using group_vars instead, you can add a host to multiple groups. Then deploying the same services on a new hosts would simply be adding it to the group

Just make one super strong password, use that to unlock you password manager and have it generate 30 character passwords for everything.

I’m sure we could make a graph how less health can translate to better “toughness” through higher AC

It’s probably just easier to use public certs with DNS verification than building and distributing your own certs.

Funny, I switched from GUI to CLI years ago because that was more reliable for me

I don’t get it, are GUI updates faster or something?

I have some connected with ZigBee to my Home Assistant. Their new “Essential” line is price competitive with most other smart bulbs you can find and they’re working great so far.

It probably is accidental, but they don’t care enough to fix the root problem

Making the term self hosting exclusive to running stuff at home feels unnecessarily elitist. Not everyone has the space, bandwidth or family approval to run stuff at home.
You can have the term homelab if you want

Thanks for the explanation, it sounds very interesting.

I currently use something else for podcasts, but better mobile apps are a big draw.

I live in a country where making copies of movies and having them for private consumption isn’t illegal.

I wouldn’t blame the Jellyfin devs for this situation, they inherited a lot of bad code from Emby and are still cleaning it up.

I don’t have an issue with Plex. I don’t use it

Most of these require some form of random id to exploit, which leaves you either brute forcing ids or brute forcing a user account

I think Amazon and Qobuz both offer some music for purchase.