I have 8 active in Manila, all game servers. 6 for actual games, 2 that nobody has joined in 4 months but I keep them up because the configs took me a weekend to get right. Latency to Singapore is 45ms, to Tokyo 65ms. I measured this myself.
I could kill those 2. The 6? My imaginary friends would notice.
I meant the cold spare question. Do you firewall something that has no traffic? My 3 spares are literally powered down. What is there to firewall?
Member
Thirty client sites, spread across three OVHcloud boxes in Gravelines and two in Amsterdam at a local colo. I could shut down zero without phone calls. The clients notice before the monitoring does.
I do not count the test box I keep forgetting at Hetzner. It has been running since March.
Powered down is not cold spare, that is archive. Cold spare means ready to swap. If it boots and has your ssh keys, it gets a firewall. Mine run nftables with drop-all default, I activate the ruleset when I bring them online.
The trial thing is real though. I found a DigitalOcean droplet from 2021 last month. $6/month for a metrics experiment I abandoned.
I run 34 WordPress instances on 2 servers. Could I shut down one? Sure. Will I? No, because the day I do is the day the remaining one has a kernel panic. I have seen this movie.
Member
This happened. Not to me, to a colleague. OVHcloud, 2019, they will not name the datacenter but it was not Roubaix. He saw ARP traffic from another tenant for eleven hours. He reported it, they fixed it, he moved to dedicated.
Regulation says report within 72 hours. He reported in 2. They were not grateful.
Member
Total beginner here. I have 1 server. It runs pihole and a minecraft server for my nephew. I do not know if it has a firewall. How do I check?
Member
sudo ufw status
Or if you are on Debian with no ufw, sudo iptables -L
If you see mostly ACCEPT rules and no DROP at the end, you are naked. Fix that before the nephew invites friends.
I have 5 active in Auckland, 2 in Sydney at Vultr for AU clients. Plus 3 "cold" spares that I test restore to quarterly. Last test, two of three failed to boot because I had updated the base image and forgot to update the spare config.
3-2-1 is backup. Test your restore. Test your spare. I am boring at parties.
Member
Small services for local clients in HCMC. 6 servers, all in-country at a local provider. Latency to Singapore is 35ms, to Tokyo 70ms, to US west coast 180ms. My clients do not care about US latency, their customers are here.
I could shut down 1, it is a staging server that somehow became production because a client "needed it live for a demo" in 2022.
