Or just use mosh. I learned this running nested Proxmox on a Contabo VPS in Singapore because I hate money and sleep. The latency from Bangkok to their Singapore DC with plain ssh was death by a thousand drops.
Mosh doesn't care. Mosh survives.
Or just use mosh. I learned this running nested Proxmox on a Contabo VPS in Singapore because I hate money and sleep. The latency from Bangkok to their Singapore DC with plain ssh was death by a thousand drops.
Mosh doesn't care. Mosh survives.
You absolute maniac. I love it.
My OVHcloud tunnel was to their London DC from Bristol, so latency was fine. But I did have to open a ticket because the rescue mount filled up. Their support was... there. Eventually. Took 18 hours for a "please clear some space" reply.
If you're going the OVHcloud route, size up. Double what you think.
Update: Mr Whiskers has now claimed the warm spot where the ARM box used to sit. I think hes mocking me.
I got about 400GB off so far. Plex metadata is the worst part, its all these tiny files. rsync --progress looks like its having a stroke.
And yeah I know I know, 3 years upfront. I was trying to be financially responsible. Turns out "sunk cost fallacy" applies to cloud too.
Windows Server Backup would handle this better. I said what I said.
Actually no it wouldn't, I'm just bitter that my Vultr instance running 2019 Standard is still somehow more reliable than this ARM story. Four years though. That's a good run for cheap hardware.
Anyone know if Contabo is keeping the x86 VPS line? Asking for a friend who also made poor life choices.
Try -P instead of --progress, and add --info=progress2 for aggregate speed. Tiny files kill throughput because the per-file overhead dominates.
I benchmarked this last month moving 2M files between Hetzner regions. Went from 2 hours to 18 minutes with proper flags. Numbers over vibes.
Salam kenal all. I run small hosting from Jakarta, mostly resell Hetzner and OVHcloud to local clients.
This Contabo news already spread to my WA group. Three clients ask me "our VPS safe or not" because they see cheap ARM and want. I tell them: I don't sell Contabo. Not because bad, but because I cannot explain "your server maybe gone in 72 hours" to client who pay yearly.
Monthly only. I sleep better.
You were trying to not think about it for three years, you know. We all do this, you know. The subscription is the modern ritual, you know. Pay yearly so you don't feel the monthly death, you know.
Then the death comes all at once, you know.
Carl has entered his "we all die, you know" era again. Someone get him a juice box.
I'm in Bogotá and I was literally about to buy one of these ARM boxes for a project. The email timing saved me, I guess? But now I'm looking at Hetzner's CPX11 in Helsinki and the latency to South America is... not great.
OVHcloud has nothing in Latin America except São Paulo which is still far. Vultr has São Paulo too. Options are thin down here unless you want to pay local premium prices.
What are other people in this region doing?
KL here. Same problem, different ocean. I ended up on Hetzner Singapore even though the price is EUR and my accounting hates me. The Contabo Singapore DC was actually decent for latency to Malaysia, but if they're killing ARM maybe they're not long for that market either.
OVHcloud Singapore exists but their APAC routing has been weird since 2023. YMMV.