RackNerd took 4 days to reply to my last ticket and just said "node is within acceptable parameters." I moved my Plex to a Contabo VPS, yes I know the reputation, but at least their disk is honest about being slow.
Contabo honest lmao. Their "NVMe" is 200 MB/s on a good day because they put 40 people on one Samsung drive. But you know that going in, it's in the reviews. RackNerd markets differently.
Member OP
Day 3 here. Got the auto-response about high volume. My crew's already scattered to Valorant anyway, maybe this is the sign to let the old box rest.
Following. I have the same RackNerd plan in LA, haven't benchmarked yet because it's just a wireguard bounce. Will run YABS tonight and report back.
In Finland we have UpCloud which is expensive but you get what you pay for. I used to chase these $1 VPS deals, now I value my time more. 45 MB/s means you're spending hours debugging something that isn't your fault.
Member
Hetzner won't take my Colombian card without SEPA, and they want passport scan. I self-host at home on a ThinkCentre with 1G symmetric from Claro, beats dealing with KYC for a server.
This. The race to bottom on VPS pricing broke the whole market. Real NVMe costs money, nodes cost money, support costs money. Someone paying $10/year is the product, not the customer.
It's been two days, what'd you get?
312 MB/s seq write. Not amazing but usable. So LA node is fine, glasgow_tom probably got a bad Chicago or NY deployment. Still not "blazing fast NVMe" but at least it's SSD-tier.
Same here. RackNerd Dallas, 280 MB/s. I run a small gmod server, no tickrate issues. 45 MB/s is a bad node or ZFS send happening, not the whole product line.
