Winding Down ApertaTube
We are winding down our current PeerTube instance and replacing it with a new one. Here's why.
The time has come to make the difficult decision to wind down our PeerTube instance, ApertaTube.net.
We will instead be replacing it with our new instance, TheNewOil.tv.
In this blog post, I will explain the original vision behind ApertaTube, the decisions that led to its shutdown and replacement, and the differences between the two.
The Vision
PeerTube has several things working against it when it comes to getting a strong user base and mainstream adoption, especially compared to other Fediverse services like Mastodon or PixelFed.
First: the UI is a little dated and content discovery on other instances is extremely difficult. (I assume the apps also have some catching up to do, but I've never used them so I don't know.)
Second, PeerTube has a content problem. In addition to simply not having a lot of content in the first place, most PeerTube instances fall into one of a few categories:
- Far-right hate speech
- Someone's personal repository of random clips
- Highly-specific, niche content (such as urbanism or gardening)
- Abandoned
- Does not allow uploads
Third, there's the technical challenges, which I'll get to a little later.
In my personal opinion, these all combine to make PeerTube more of a statement than actual useful ecosystem:
- The UI and lack of discovery make it hard for end-users to enjoy their experience.
- The state of instances deters creators from joining PeerTube, since they pretty much have to create their own.
- The cost of creating your instance (in both tech skills and money) is pretty much the nail in the coffin.
My original vision for ApertaTube was a service similar to TILvids.
TILvids is a PeerTube instance that allows uploads only for approved creators, aiming to create a high-quality, curated experience instead of simply opening up the floodgates and allowing everything (like YouTube and others).
My main complaint with TILvids is they don't federate with other instances. They act as a walled garden. I find this antithetical to the spirit of PeerTube, the Fediverse, and open source in general.
I wanted to make something more open.
My original vision was that ApertaTube would eventually become an open platform - independent of The New Oil - with their own funding, moderation staff, and creators uploading.
My plan was to sustain ourselves through donations, or at very least through paying subscribers who would pay a small amount in return for storage space, perhaps on a Freemium model.
Unfortunately, both the interest and income streams failed to materialize.
Over the years, I only had a few people reach out to me to ask for upload permissions, and almost all those who did couldn't afford or didn't want to pay. They wanted to upload just a few videos for personal reasons, such as backup or sharing with friends and family.
I am no stranger to the fact that things often take time to take off, but I ran ApertaTube for over four years. It's clear that it simply isn't growing the way I had hoped.
The Technical Challenges
Running any public service can be a challenge in several ways.
- Technical: Some services (like Mastodon) have incredible documentation and very easy installation processes. Others (like Matrix) expect you to have a high degree of technical skill and do very little hand-holding of the process. (Not to mention your pre-existing knowledge of how to get a domain name, set up DNS/Lets Encrypt, etc.)
- Legal: If someone uploads hate speech or child sexual abuse material, I can potentially be liable for that if I don't handle it correctly. This is even worse for cloud providers or other types of services where users may not find and report the material, so you could easily miss it.
PeerTube in particular faces an additional challenge that services like Mastodon don't: video processing and storage is hardware-intensive and expensive.
According to PeerTube's official documentation, the bare minimum specs for a server are 1 CPU core, 1.5 GB of RAM, and "enough storage for videos" but the ideal situation would be at least 8 cores and 8 GB RAM.
On the lower end of the scale, this is pretty affordable at less than $10/month for a VPS from Hetzner or Namecheap, but you pay the price in performance: videos often fail to transcode, take several hours when they do, and the site can be slow to navigate.
An actual 8/8 sever starts at $20/month from Hetzner but quickly climbs into the triple digits from other providers.
But the real killer is storage space.
Storage
It's hard to nail down "average" file sizes because there's so many factors, but I'll do my best.
A 10-minute HD video (1080p) can range anywhere from 300 MB to 1.8 GB depending on bitrate. For streaming quality (8 Mbps) it can be as little as 600 MB, while high-quality (10-24 Mbps) can be 750 MB - 1.8 GB.
For context, the maximum file size for a photo on Pixelfed is 15 MB, while an all-text post on Mastodon (500 characters) would be less than 2KB.
In other words: in the same space it takes to host a single streaming-quality 10-minute video, you could host 40 PixelFed photos or 300,000 all-text Mastodon posts.
When you run a public PeerTube instance, you should cater to multiple connection speeds and bandwidths - some people may be on older devices or slower internet. That means uploading multiple versions of the same video in different resolutions.
At a bare minimum, I (and most people) would recommend 1080, 720, 360, and maybe audio only. This means every 10-minute video you upload takes up just over 1 GB of disk space.
Depending on your hosting provider and which VPS package you choose, your install can have anywhere from 20 to 160 GB of space.
This means that on the low end, you can store just under 20 non-4k videos (when you account for variations in length, captions, thumbnails, and the PeerTube software itself).
For an individual content creator, that's not unreasonable, especially at $5-10/month. But for a public instance, things quickly fall aparts.
For starters, we would need to upgrade hardware - which isn't unreasonable. A 4 CPU/8 GB VPS from Hetzner is only about $10/month, and that comes with 80 GB of space. But again, that's only about 80 videos.
In the past 12 months, not including shorts:
- Privacy Guides has uploaded about 35 videos
- All Things Secured has uploaded about 40 videos
- The Linux Experiment has uploaded about 90 videos
- Techlore has uploaded about 130 videos
It's also worth noting that many of these videos are much more than 10 minutes - some up to or over an hour.
A small VPS's included storage is plenty for a single creator, or a handful of part-time creators. But for a lot of creators uploading videos regularly, storage would very quickly become a problem.
The Failed Migration
Last month, my annual plan with 1984 Hosting was up for renewal. The plan was to switch to a different hosting provider and setup that was more scalable, powerful, and affordable.
ApertaTube is currently hosted on a #2 VPS (1 CPU, 2GB, 50 GB storage) + 100 GB storage space upgrade. This setup runs us about $50/month.
The intended destination was Hetzner, where a comparable VPS (2 CPU, 4 GB RAM, 40 GB storage) was only about $7/month. The additional storage problem would be solved with an S3 bucket (also from Hetzner) starting at 1TB at an additional $8/month, offering me a more powerful server with exponentially more disk space for less than half my current cost.
This would also give us more scalability: since we were hosting content on an S3 bucket, we could simply keep adding storage space as-needed without buying increasingly powerful servers simply to expand our storage capacity.
This is where it all went wrong.
PeerTube's migration documents are relatively straightforward, and for the most part I was able to navigate them with only a little bit of help from others.
The challenge is that PeerTube is not designed to go straight from "everything on one VPS" to "move PeerTube itself and then move all the files to S3 separately." The proper procedure is either to move everything to the new VPS and then to S3 or move the files to S3 first then move the remaining files to the new VPS.
Long story short: this was all information I discovered in-progress through a lot of trial and error. At one point I had two servers running simultaneously. Things became out of sync, there were errors in the setup, and there just too many moving parts for me to be accurately chase down what was failing where.
The Aftermath
At the end of the day, I was forced to step back and take stock of things:
- ApertaTube wasn't growing the way I wanted.
- 1984 was simply too expensive to keep using.
- The migration had become horribly mangled.
It simply didn't make sense to keep fighting to untangle things.
So in the end, I've decided to sunset ApertaTube. I will keep it running through the end of November, at least. I'll try to do through the end of the year. This is so that users have time to migrate their accounts to other instances if they wish.
TheNewOil.tv
However, I have no intention of abandoning PeerTube, the Fediverse, or the open source ethos. I want a place where people can view my content free of trackers and Big Tech, where I'm not at risk of being de-platformed, censored, or having my audience forcibly taken from me.
I have already finished setting up a new instance: TheNewOil.tv. This will serve as a dedicated place for The New Oil's content. They're all there already now.
Currently we are using the CX23 + S3 setup I mentioned before.
If a CX33 (4 CPU, 8 GM RAM) opens up, I plan to upgrade immediately and transfer files away from S3. This would be fit the recommended minimum performance according to PeerTube's documentation (4 CPU, 4 GM RAM), with 80 GB of storage space on the same disk.
As I said before, this setup would be plenty for a single creator, especially one putting out content at my rate (even at my ideal rate of 1-2 videos/month, I don't publish in 4k so that would take me somewhere in the ballpark of 5 years to get even close to filling up the disk).
And if I do need to return to using S3 someday, this new setup is already pre-positioned to easily transfer back without the issues we suffered before.
While this new setup is much less expensive, we still depend entirely on audience contributions to keep the bills paid and blinky lights blinking on our servers. If you get value out of our work, please consider signing up for early access to content and more!
ApertaTube vs TheNewOil.tv
So to recap, here's the TL;DR:
ApertaTube will shut down no earlier than November 30, 2026 and no later than Dec 31, 2026.
TheNewOil.tv will not attempt to be an instance for multiple creators. It is strictly for TNO videos on the Fediverse.
That said, we still don't want to become an insular walled garden. We still allow follow requests, and we will continue to follow other instances we feel are interesting and high-quality.
We also allow public-signups if you feel that we've treated your data well and would like to continue using our instances.
I apologize to anyone who felt like ApertaTube was home. I hope you'll find a new home that feels just as cozy, whether with us or somewhere else. I also apologize that we weren't able to create another major space in the Fediverse. Maybe someday in the future when we have more resources.
If you appreciate the work we do, please consider supporting us financially so we can continue to host these services and not have to rely on centralized platforms.
Tech changes fast, so be sure to check out our website for all the latest recommendations, tools, services, and more.
