I haven’t looked in a few years but 20TB is probably plenty. I agree that Wikipedia lost its way once it got all that attention online and all that search traffic. Everyone should have their own copy of Wikipedia. I used to download the daily incremental data dumps but got tired of it. I still have a few TB of them around that I’ve been wanting to merge.

The text is in not-exactly-convenient database dumps (see other commenter’s link) and there are daily diffs (mostly bot noise), but then there are the images and other media, which are way up in the terabytes by now. There are some docs, maybe out of date, about how to run the software yourself. It’s written in PHP and it’s big and complicated.

Changing employment status is a qualifying event or whatever it’s called, so assuming the company offers its employees coverage immediately on hiring (not all do), you shouldn’t have to wait for open enrollment. However, while health coverage is a common benefit of employment, it’s not universal and policies vary by employer. That is: ask the company, or if for some reason you don’t want to ask, get a copy of the employee handbook, maybe by asking one of your co-workers for it. The info for the particular company is likely to be in there.

How much do you expect to pay for the 24 NVMe disks?

It’s possible for a while but there is a whack-a-mole game if you’re doing anything they would care about. So you will have to keep moving it around. VPS forums will have some info.

Never used any of them and get even more glad of it with passing time.

I’m still driving mine and would be very reluctant to swap it for a modern enshittified car. I sometimes think of homebrewing my next car (DIY EV conversion of an older ICE car) rather than put up with any manufacturer’s offerings. Who knows.

Noo, really, idk what Disco was but tags and recommendations from other humans are plenty to find good AO3 fic to read. And AO3 itself has been getting hammered for months, presumably by corporate AI crawlers. A recommendation engine would also have to crawl AO3. That’s very difficult to do because of said hammering. Even the regular download feature barely works now if you use fanficfare for it.

Are you familiar with git hooks? See

https://git-scm.com/book/en/v2/Customizing-Git-Git-Hooks

Scroll to the part about server side hooks. The idea is to automatically propagate updates when you receive them. So git-level replication instead of rsync.

I see, fair enough. Replication is never instantaneous, so do you have definite bounds on how much latency you’ll accept? Do you really want independent git servers online? Most HA systems have a primary and a failover, so users only see one server. If you want to use Ceph, in practice all servers would be in the same DC. Is that ok?

I think I’d look in one of the many git books out there to see what they say about replication schemes. This sounds like something that must have been done before.

Why do you want 5 git servers instead of, say, 2? Are you after something more than high availability? Are you trying to run something like GitHub where some repos might have stupendous concurrent read traffic? What about update traffic?

What happens if the servers sometimes get out of sync for 0.5 sec or whatever, as long as each is in a consistent state at all times?

Anyway my first idea isn’t rsync, but rather, use update hooks to replicate pushes to the other servers, so the updates will still look atomic to clients. Alternatively, use a replicated file system under Ceph or the like, so you can quickly migrate failed servers. That’s a standard cloud hosting setup.

What real world workload do you have, that appeared suddenly enough that your devs couldn’t stay in top of it, and you find yourself seeking advice from us relatively clueless dweebs on Lemmy? It’s not a problem most git users deal with. Git is pretty fast and most users are ok with a single server and a backup.

I wonder if you could use HAProxy for that. It’s usually used with web servers. This is a pretty surprising request though, since git is pretty fast. Do you have an actual real world workload that needs such a setup? Otherwise why not just have a normal setup with one server being mirrored, and a failover IP as lots of VPS hosts can supply?

And, can you use round robin DNS instead of a load balancer?

What does this even mean? You want to replicate between git repositories? Can you do that with receive/update hooks on the servers?

Yeah a buddy of mine (not rich) has one and seems to like it. It’s the big style not the Moto Razr style. It’s like two normal sized smartphones folded together, so when you open it you get a big roughly square screen about 6 inches on a side. About 2x the area of a normal phone screen. It’s a Samsung, idk what model or what it cost. It looks nice. No idea about fragility. If you have a question I can relay it to him.

You can get that much ram for 5k but maybe not in the densest modules. So you need a mobo with 2x as many socke5, raising costs again. I might look into it but I have trouble thinking of applications that can use so much ram. 256gb is quite easy to do and is also a lot.

Last I saw you had to use super premium ultra dense memory modules to get 1tb into a motherboard. Maybe that’s less so now. But the hope would be to use commodity ram and CPUs etc. 10k for a 1tb system is pretty good. Last I looked it was a lot more.

There are few motherboards with enough dram channels to handle a TB of dram. That’s basically why ram drives existed back in the day, and they are still potentially sort of relevant for the same reason. No a TB of ram wouldn’t fit on an m.2 card, but a box of ram with a pcie connector is still an interesting idea. Optane also aimed for this, but it never got enough traction to be viable, so it was scrapped.

Added: also look at kimsufi.com as an alternative to Hetzner.

Dedi will perform a lot better and be more consistent and reliable. They’re not THAT expensive if you’re making nontrivial use of them. Otherwise maybe you can keep moving around between Contabo products. Keep in mind too that hdd performance will seem a lot better when you’re not sharing it with dozens of other users. I have an HDD server and it’s fine for browsing. Might not be great for large seek-intensive databases but I’m not currently doing that

Anyway you can also ask on lowendspirit.com which is a forum about budget vps.

Get a Hetzner dedicated server. Don’t mess with vps once it gets that large. Look on hetzner.com/sb for auction servers that might be better deals than regular ones. And does your 800gb of storage really have to be all SSD? Hard disks still exist and cost lots less.