Things on infosec.exchange were getting a little bit sluggish, and I've had complaints about media processing taking a long time so I added another app server for the US region. In the past, that has caused media uploads for some people to just fail outright because the reverse proxy hops the upload from one app server to another - and the other has no idea what to do with the last half of a file. I am hopeful I've got that problem ironed out in the nginx config, but would be interested in knowing if anyone runs into that...
@jerry@webhat@dustinfinn have you tried wasabi? It’s a bit cheaper than backblaze (with big catch that you have to pay for 60ish days of storage, if object was removed earlier than 60ish days) and they have multiple regions
@jerry@webhat@dustinfinn I’ll be happy to help, if you need anything!I know few places where you can find vps with decent bandwidth for a few bucks a month
@jerry@webhat@dustinfinn 250$ a month per server. Not bad. What kind of config you need? 250 on a hetzner is kinda beefy? Like W-2295, 128G RAM & 2x 1TB NVMe?
I also ran into a really odd iptables problem... when I meshed in the second app server node using wireguard, my egress filter rule was blocking outbound traffic on both hosts. After some fiddling, out of frustration, I flushed the rules out of iptables on both hosts - so no rules. And... iptables was still blocking the outbound traffic. I ended up having to reboot each host - and once I did that, things worked ok. Has anyone seen iptables go into zombie mode before?