Skip to content

OVH fire anniversary, who's still there?

VPS Hosting by sarahjey 26 replies 1.4K views
#1

Sorry if this is a dumb question. I was reading old threads and saw the big fire outage from years back. Curious if anyone here was affected and what they learned? I only started with VPS last year so I missed it.

I'm on a super cheap plan with some random provider and now I'm worried. Do you all actually do offsite backups or just hope for the best? Thanks everyone!

learning on $5 VPS and prayers
#2

Was on server in Strasbourg. Total loss. I think I had backups but never tested. Thanks to that fire I learned hard way. Now I run 3 copies on different providers. Vultr, Hetzner, and one local NAS. Fire changed my mind about "the cloud" being safe place. It is just someone else burning building.

#3

Still here lor. Moved to CloudCone after. Can. Already setup proper backup before but never verify. Fire day I find out my "automated" job was empty for six months. Siao liao. Now I check leh, every month, manual download restore. Kiasu but better than cry.

9 #4

I monitored 14 services during that event. Uptime dropped from 99.97% annual to 73.4% in a single afternoon. Status pages updated 4 hours after smoke was visible on social media. Alert fatigue is real, but that day my PagerDuty saved actual data. Post-incident, I implemented:

  • Geographic redundancy minimum 300km apart
  • Quarterly restore drills with checksum verification
  • Status page polling every 60 seconds, not 5 minutes
Three providers admitted they had no tested recovery plan. One never recovered customer data.

436 days. reboot is surrender.
#5

Observed packet loss exceeded 10^2 percent during the incident window. Observed recovery time exceeded 10^4 seconds for affected prefixes. The provider's published SLA allowed compensation of 10^1 percent monthly credit per 10^3 seconds downtime, if I remember their terms right.

RFC 2119 "REQUIRED" does not apply to physical disaster clauses in most TOS documents[2]. I observed three peers who assumed "cloud" implied geographic redundancy by default. This assumption is not supported by measurement.

[2] RFC 2119, Key words for use in RFCs

100% packet loss at hop 3
7 #6

Just use RAID-1 across datacenters

Skill issue if you trusted one provider with no exit plan. The fire was 2021. We are in 2025. If you are STILL single-homed, you deserve what happens. Just use rsync + zfs send to a second cheap box on another continent. Costs maybe $7/month. Cheaper than one hour of your time rewriting lost configs.

Also: "the cloud" is just a computer you cannot see burning. Poetic. Stolen from someone on fedi.

oops: 0000 [#1] SMP
1 #7

How do I actually test a restore? I run a small shop on Hostinger and I download my cPanel backup sometimes but I've never tried putting it anywhere else. I'm worried I'll break something if I test it wrong.

Is there a safe way? I don't know much about servers, I just sell pottery online.

frames, tables, still valid HTML
#8

I had the same empty backup job with my old host

#9

Three providers sounds expensive lor, what's the monthly hit like

#10

Which Hostinger plan, shared or reseller

Post a reply

You need an account to reply. Log in or register to join the conversation.

Post reply Preview Save draft