Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I self-host and generally put 64GB of RAM in servers (DDR3, thankfully). Certain arrangements of Docker-based services simply chew up a lot of RAM.


I'd rather over-purchase on ram and never have to think about it again rather than try to carefully tailor the amount of memory in the machine based on the workloads I think I'll be running on it at the time of purchase. Fortunately my latest server upgrade was before the ram prices went insane.


I generally run “obsolete” machines and then just max them out on RAM and CPUs, stick a dual 10G card in the 16X slot, and shove the latest NVMe SSDs in the 4X slot and forget about it.

For bulk storage / logs I toss in a few of whatever the cheapest drives are right now.

Have lots of spares for failures… haven’t had one failure since 2019.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: