ZFS replication to another auction box. The "no SLA" becomes "self-SLA." I keep a cold standby in Nuremberg, replicate from Falkenstein nightly. If primary dies, I boot the standby, update DNS, done.
Total cost: €50/month for two boxes. One provider, no SLA, but I have my own continuity.
Now THIS is how you ZFS. Not on a VM. Not for compression alone. Actual disaster recovery with send/recv and a cold standby.
I retract my earlier criticism. You are forgiven. RAIDZ calculations available upon request.
I am good, thank you. My arrays are mirrors. Simple, fast, no parity calculation anxiety.
The thread has truly died and been reborn as a ZFS thread. We should stop.
One more negotiation data point before we let this become a filesystem cult.
Tried the "I'm leaving" technique with Vultr. Had a $40/month instance for 18 months. Told them I was moving to Hetzner for price. Their reply: "We value your feedback. Your cancellation has been processed."
No retention offer. Nothing. Vultr genuinely does not negotiate. Confirmed.
Member
This matches my experience and everyone else's. Vultr's model is API-first, no-touch, thin margins. They don't have a retention team because they don't have the margin for one.
Their value is instant provisioning and global locations. You pay for that convenience, not for relationship.
Exactly. Vultr is for when you need a Mumbai instance in 60 seconds for a test. Hetzner is for when you know what you need and can wait a day for setup.
Different tools. I use both, just don't waste time negotiating with the wrong one.
Member
Wellington, NZ here. Mostly lurking, occasionally useful.
Can confirm Vultr's Sydney to Wellington is 32ms, their Auckland location (if they had one) would be better but they don't. I use their Sydney for latency-sensitive stuff, US locations for price-sensitive.
Has anyone mentioned that Vultr's Melbourne location is actually more expensive than Sydney for the same plan? Found that out the hard way on https://www.vultr.com/pricing/. No idea why.
Member
Wait, what? I thought they were the same price. Let me check... you're right. $6 vs $5 for the smallest plan. That's bizarre given Sydney's higher datacenter costs generally.
Maybe Melbourne is newer and they're amortizing build cost? Or just testing price sensitivity? Either way, good catch.
Vultr's Melbourne opened in 2020, Sydney in 2019. So not newer. More likely capacity management - Melbourne has fewer nodes, higher demand ratio, so they price-discourage new signups.
This is common with cloud providers. The price is a load balancer.
Following. Good thread.
Has anyone tried negotiating with HostHatch specifically? I like their storage VPS locations but the regular compute pricing seems high compared to RackNerd. Wondering if they match or come close.
