Hacker News Reader

Comment by andrewaylett | Hacker News Reader

The reason I'm considering a Pi cluster is resilience and repeatability. The reason I don't have one yet is because (like you) I'm unconvinced it's the right way to get that.

At least in theory, a Pi cluster has better failure modes than a single machine even if it's less powerful overall. And yes, I'm currently running on an old laptop -- but it's all a bit ad-hoc and I really want something a bit more uniform, ideally with at least some of the affordances I'm enjoying when deploying stuff professionally.

View on HN

Your house a single failure domain, you're not really going to be resilient to a lot of common failures, and most of these home labs have every device plugged into the same UPS, so there's really no difference between the power failure domain of 10 Pis and one desktop computer plugged into a UPS connected to your home router.

Do yourself a favor and buy a NAS and a compatible UPS. Any modern NAS software will speak one of the UPS IP protocols to handle graceful shutdown if your power goes off. Once you have the money, buy a second NAS and put it in a relatives house, set up a wireguard/tailscale tunnel between the two devices, and use it for offsite backups.

Separating storage, compute, and database into dedicated appliances makes sense for home labs. Getting redundant units early is key.

A pre-built NAS (configured to raid 5 at least) is worth the cost - storage should be set-and-forget since drives will fail and hot-swapping drives and automatic rebuilds should be zero downtime to life. Commercial NAS solutions have proven backup workflows.

For compute and databases, home setups can mirror to cloud or remote locations. Proxmox makes this straightforward with its web admin - just a few clicks to spin up replicas.

Modern consumer hardware and internet are quite reliable now. Business-grade equipment performs even better.

It's not power failure I'm worried about, it's failure of other systems. But even so, yes, it's more likely that the systems that are supposed to ensure redundancy will fail than the otherwise-non-redundant services.

For what it's worth, my house isn't a single power domain -- I'm running some of my services on an old laptop that's still got some battery. But it goes to sleep if the power goes off, and needs physical intervention to recover. Which is an excellent example of redundant systems providing multiple single points of failure and of the perils of using random left-over hardware to run stuff.

Some people suggest the thin clients as an option. You can buy them used and they're x86, usually expandable RAM and storage.

Reply on news.ycombinator.com