kenji3
Member
OP
- Joined:
- Jul 2024
- Posts:
- 138
- From:
- Osaka, JP
I may be wrong, but I believe we have not had an honest community reckoning since the fire. I moved my backups to APAC after that event. Singapore and Tokyo regions.
How many here actually changed their practices? Or was it just two weeks of panic and back to normal?
I am curious if trust in European providers recovered.
conbini > datacenter snacks
kate3
Member
- Joined:
- Jul 2024
- Posts:
- 208
- From:
- Utrecht, NL
How many here actually changed their practices?
The virtualization tax on proper redundancy is real. After the fire I stopped trusting single-location anything. OpenVZ containers with cgroup limits made migration impossible for some workloads. KVM with proper qcow2 snapshots saved me.
I now run replicated storage across three regions. Latency hit is acceptable. The kernel-level isolation matters when you are evacuating hosts.
virsh list --all | wc -l: 47
kate3
Member
- Joined:
- Jul 2024
- Posts:
- 208
- From:
- Utrecht, NL
Singapore and Tokyo regions
APAC latency from Utrecht is brutal for interactive stuff. I tried Singapore for object storage and got 280ms. Fine for backups, unusable for live databases.
I ended up with Hetzner Helsinki plus OVH Limburg as my pair. EU-wide, sub-30ms between them.
virsh list --all | wc -l: 47