I am a software developer by craft and a linux system admin by hobby. I cannot commit to moderating and managing my own instance, but I would be glad to help someone with the technical aspects.
The most common complaint I saw in Reddit and here about switching to Lemmy is the difficulty of setting it up, so I thought I would help bridge this gap.
While I have never hosted my own instance before, I already checked the setup guide and it looks pretty simple to me, so I am confident I can do it. Please feel free to comment or DM.
It would be great if you can comment general questions. I can then respond to you here and maybe others will see it and know how to host their own instances too.
An up to date step by step instruction for docker, ansible, or scratch setup would be a fantastic step in the right direction. I’m not a complete moron, but I feel the current guides are lacking. There’s a lot of assumptions on the admins knowledge of the associated systems beforehand. As someone else said, the first to make a turnkey solution… hot dog. Would absolutely blast adoption into space.
What was it you were struggling with? I found the guide to be straightforward, but I also have a tech background similar to OP
I likewise have a similar background to the OP. I wouldn’t say I struggled (thanks to the Matrix chat room), but the docker instructions straight up didn’t work a couple of days ago. I and another guys submitted a couple of PRs and I think the instructions should at least work for common setups now, but the edges are very rough.
By far the best way to contribute IMO is submitting PRs. The devs seem very willing to take contributions without making you jump through hoops.
Thanks for offering your knowledge! I successfully set up and instance using their docker installation guide. However I was never able to get the smtp server to work. I first tried to add postfix to the docker-compose file like they have in the ansible installation example on github, but that didn’t work. Just trying to add an email address to my account would stall the UI with a spinning animation on the Save button. I then tried to update the hjson config file by adding my sendgrid api credentials and removing postfix from docker compose. That gave me the same result. At that point I kinda gave up and deleted my vps. I don’t have access to my error logs anymore, but I can spin up a new vps to try to get the same errors again if needed.
A lot of VPS providers block port 25 (and other email ports) because they don’t want people to set up bot spam mail servers on their services. Could that be the issue?
I’m using a VPS with Hetzner. I’ll have to look up if it’s blocked. Thanks for the info.
I’m in the same boat and confirm that yes, port 25 is blocked.
Port 25 is almost universally blocked because of this. It’s also blocked on a lot of residential ISPs for the same reason. This is probably one of those things where the documentation could be improved. A quick Google of the docker container that is being used for postfix says that we should be able to use another SMTP relay through gmail or another provider as an example, although I have not tried yet.
If you have Cloudflare you can set up an email alias for incoming email and then create a secondary Gmail address on top of your existing one for outgoing email. If you go to ‘Settings > Accounts and Import > Send mail as’ and add another email address (not an alias) with the same email as the one you setup on Cloudflare (noreply@yourdomain.com). You will likely need to create a Gmail app password to sign into the email server if you use 2FA.
Once you’ve created this email Gmail will send you a confirmation email to confirm it’s all working. Then you can just enter Google’s SMTP server info for Lemmy along with your email (your actual email, not the CF one), and the app password you created.
If done it this way for a few services beyond Lemmy and it’s worked well so far. This way you’re also using a Gmail account technically so you can hopefully avoid blacklists and spam filters.
Amazon has a very generous free tier for outgoing email in SES, and it is pretty easy to set up.
Do you think it would be possible to host using a raspberry pi 4?
How much traffic does Lemmy create daily?
CPU requirements for Lemmy hosting are minimal. Memory is useful - you’d want to use the Pi 4 with either the 4GB or 8GB RAM, anything less than that will work but you’ll be running the risk of difficulties if the server gets busy.
You’ll also need plenty of storage, especially if people are going to start uploading media to your Lemmy host. Given that a Pi runs off an SD card you might well find yourself running out of storage space - I’d recommend attaching a USB storage device for the reassurance in that respect.
I think for a small number of users, yes.
I run lemmy.studio on a VPS with 1GB of ram and 1 VCPU, so a raspi4 should suffice, at least initially. Bandwidth is around 7.5 Mbps.
Perhaps write a thread, with step-by-step text and photos/screenshots per platform? And then the next one to do would be a step-by-step video?
Example, I’m not technically minded, but have an Unraid box with a 1Gb connection. There are extra steps there that I’m aware of but have no idea how to do because it’s over my head.
A more detailed set up guide would be great! I feel like some details were missing from the Lemmy guide.
Here’s an example of how to build it from scratch. https://lemmy.grouchysysadmin.com/comment/7413
A couple of us have been detailing experience and problems with those instructions:
!lemmy_admin is the official place for discussing install and being a server administrator
Depending on how much Lemmy get’s hugged to death I might take you up on this. Thanks!
Something that would help a lot of selfhoster types would be prebuild docker images and a good example docker compose. (something kbin could also use)
I have made a request at linuxserver.io.
https://discourse.linuxserver.io/t/kbin-and-lemmy-fediverse-reddit-alternatives/7667
I feel like there’s a lot of money on the table for the first person to set up a turn-key hosting platform for Lemmy. Something like MoltenHosting but for Lemmy instead of Foundry.
There’s similar things for mastodon and a project currently in alpha/beta for multiple platforms, incl Lemmy, called spacehost.
See mastodon post here https://mastodon.social/@atomicpoet/109905703584075719
See their main rep on the fediverse here: https://atomicpoet.org/users/atomicpoet and https://calckey.social/@atomicpoet (@atomicpoet@calckey.social @atomicpoet@atomicpoet.org).
They also admin calckey.social.
Hi there - thanks for posting. I’m serving lemmy (ansible install) on nginx from Ubuntu 22.04 and I’m seeing this weird problem sometimes where the submit button is not enabled on the “create post” form. Sometimes it works but then it inexplicably stubbornly refuses to - I am not seeing any apparent error show up in the console, nothing in the lemmy or nginx logs (naturally, because the HTML form isn’t submitting to the server because the submit button’s disabled). Happens in different browsers, tried restarting nginx, rebooting, nothing seems to fix it. Now, my addled brain does recall seeing something about Ubuntu 22.04 having some issues with Lemmy and perhaps that’s what I’m seeing - it’s just weird and I’ve never seen a web app do this before - any ideas? The url is quex.cc and you’re welcome to test there if you like, but I’m really just looking for any suggestions for investigation before I start thinking about moving to a different OS.
Open a bug ticket on the lemmy-ui Github. Every issue should be tracked
yes, good suggestion, thank you - although before I do I’m going to spend a bit of time to see if I can isolate the problem - I just tested it from a mobile device and the local PC browser and it’s working now. I used the exact same url, text and everything to create the post and I didn’t change anything - didn’t reboot, didn’t even restart the browser. Except for the fact that it’s intermittent, it looks like some kind of js or css problem, but I need to replicate it and if I do that, I’ll probably be able to figure it out.
I realize how stupid I’ll sound admitting this, but I think the problem was that I wasn’t selecting the “community” in the pulldown, so the submit button wasn’t enabled.
You’re awesome
I’ve been playing with my own single-user instance here using Docker. Mostly I just followed the Lemmy docs. It’s been nice & responsive and takes barely any resources, so far. I think this system can really benefit from a lot of small instances to spread the load.
The thing I struggled with the most was adapting the provided
docker-compose.yml
for my Caddy setup.I am using caddy-docker-proxy, which I absolutely love but their documentation is not the greatest for matchers.
If anyone else wants a super basic Lemmy instance running on Caddy with their domain on Cloudflare here is a
docker-compose.yml
Please make sure you update your
lemmy.hjson
hostname
field to match the domain you used in thedocker-compose.yml
for thecaddy
labelsIf you’re not using Cloudflare you can replace
build: .
(and not use theDockerfile
I provided below) in the caddy service withimage: lucaslorentz/caddy-docker-proxy:ci-alpine
(and remove thecaddy.acme_dns
label) and I believe it will fall back to Let’sEncryptversion: "3.9" services: caddy: container_name: caddy build: . depends_on: - lemmy-ui - pictrs ports: - 80:80 - 443:443 environment: CADDY_INGRESS_NETWORKS: caddy networks: - caddy volumes: - /var/run/docker.sock:/var/run/docker.sock - /opt/docker/caddy/data:/data - /opt/docker/caddy/config:/config labels: caddy.log.format: console caddy.acme_dns: cloudflare YOUR_CLOUDFLARE_API_TOKEN restart: unless-stopped lemmy: container_name: lemmy image: dessalines/lemmy:0.17.3 depends_on: - postgres - pictrs environment: RUST_LOG: "warn,lemmy_server=info,lemmy_api=info,lemmy_api_common=info,lemmy_api_crud=info,lemmy_apub=info,lemmy_db_schema=info,lemmy_db_views=info,lemmy_db_views_actor=info,lemmy_db_views_moderator=info,lemmy_routes=info,lemmy_utils=info,lemmy_websocket=info" networks: - caddy volumes: - /opt/docker/lemmy/lemmy.hjson:/config/config.hjson:ro labels: caddy: "your.domain.com" caddy.@lemmy: path_regexp ^/(api|pictrs|feeds|nodeinfo|\.well-known)/.*$ caddy.@post: method POST caddy.@accept: header Accept application/* caddy.reverse_proxy_1: "@lemmy {{upstreams 8536}}" caddy.reverse_proxy_2: "@post {{upstreams 8536}}" caddy.reverse_proxy_3: "@accept {{upstreams 8536}}" restart: unless-stopped lemmy-ui: container_name: lemmy-ui image: dessalines/lemmy-ui:0.17.3 depends_on: - lemmy environment: LEMMY_UI_LEMMY_INTERNAL_HOST: lemmy:8536 LEMMY_UI_LEMMY_EXTERNAL_HOST: localhost:1234 LEMMY_HTTPS: true networks: - caddy labels: caddy: "your.domain.com" caddy.reverse_proxy: "{{upstreams 1234}}" restart: unless-stopped pictrs: container_name: pictrs image: asonix/pictrs:0.3.1 environment: PICTRS__API_KEY: API_KEY user: 991:991 volumes: - /opt/docker/pictrs:/mnt networks: - caddy postgres: container_name: postgres image: postgres:15-alpine environment: POSTGRES_DB: lemmy POSTGRES_USER: lemmy POSTGRES_PASSWORD: password volumes: - /opt/docker/postgres:/var/lib/postgresql/data networks: - caddy restart: unless-stopped networks: caddy: external: true
Here is the Dockerfile used for the caddy container:
ARG CADDY_VERSION=2.6.4 FROM caddy:${CADDY_VERSION}-builder AS builder RUN xcaddy build \ --with github.com/lucaslorentz/caddy-docker-proxy/v2@v2.8.4 \ --with github.com/caddy-dns/cloudflare FROM caddy:${CADDY_VERSION}-alpine RUN apk add --no-cache tzdata COPY --from=builder /usr/bin/caddy /usr/bin/caddy CMD ["caddy", "docker-proxy"]
Being a highly technical guy, I struggled to get everything running on the server that I am hosting other things on. The biggest hurdle was letsencrypt, but other things weren’t working quite right. I ended up just paying for a new ubuntu vps so I can run the ansible playbook (I use arch linux everywhere else). That turned out to be super simple and “just worked”.
I was playing with the idea of spinning up a VPS through AWS lightsail since the bundled package it provides seems to be cheaper than configuring an EC2 instance/other requisite resources running 24/7.
What has been people’s experience regarding:
- Monthly data upload rates (since most cloud providers charge internet egress per gigabyte)
- Database/image storage requirements
Hosting a single user instance, I’m seeing a few GB of network traffic over the past few days and maybe 10Gb at most needed for the disk.
Hi! I am on the process to host a instance for a large subreddit to migrate to lemmy. I already played with hosting with docker before, and the ansible method seem even easier!
What I am concerned is about the storage. I need to have a way to calculate before hand because I am afraid of the costs.
I have an always free tier oracle cloud vps I could use, and buy some extra storage.
I haven’t checked but are kbin instances easier to spin up? Kbin is be able to communicate with Lemmy communities without a problem.
It’s easy from the docker from what I can tell, the bare metal / vps is a bit more as more manual pieces but nothing too crazy. I just submitted a pull request for the admin guide so that should help a bit for less head bashing… Though my instance doesn’t seem to be federating so that’s a new problem for tomorrow.