Skip to content
Register

Surviving a provider's 'datacenter migration'

General Discussion by Berlin1 8 replies 606 views
Berlin1

Member OP

#1

Contabo sent email at 11PM. "Scheduled migration tomorrow 6AM." Six[1] hours notice. Six. My production VPS, someoneelses problem now. I open ticket immediately. Autoresponder. I call. Phone tree loops. 6AM hits, server gone. DNS still pointing to old IP. New IP in control panel? Wrong IP. Another ticket. Four hours later, correct IP appears. But data from last backup? Three days old. They claim "zero dataloss migration." Their own ticket timestamps show restore completed six hours AFTER they declared "all clear." I have screenshots. I have logs. This is not a migration. This is a Datacenter disaster with MarketingSpin[2].

[1] technically five and half, I was asleep
[2] their Trademark probably

minhdock

Member

#2

1. I use Contabo too, very bad experience
2. My VPS it go down 2 day, no notice at all
3. I open ticket, they say "panned maintenance" very much
4. My backup 4 day old, I cry much
5. They offer 10% cledit, I no accept very much
6. I move to CloudCone now, better

You not alone, friend

CarlosBilby

Member

4 #3

This is caramba! My VPS it crashed in the migration too, but Hostinger, not Contabo. He took 12 hours to come back. My database he corrupt, not funny actually. Nossa, I lose the emails of customer. They give me one month free but the trust, the trust is broke. More fast I need to move, but where?

SingaporeRep

Member

Benchmark Addict

#4 
Disk Speed:
+ DD Write: 245 MB/s
+ DD Read: 312 MB/s

Network:
+ iperf3 to NYC: 89 Mbits
+ iperf3 to AMS: 34 Mbits (lol)
+ iperf3 to SGP: 12 Mbits (double lol)

Old Contabo:
+ DD Write: 412 MB/s
+ DD Read: 498 MB/s
+ iperf3 NYC: 340 Mbits

Decent disk IO, meh network. New box is downgrade. Provider won't answer why. Benchmark don't lie, support do. I used https://github.com/masonr/yet-another-bench-script

fio, iperf, geekbench. results or gtfo.
olespete

Member

Trust No One

#5

WARNINGS: migration window is EXFILTRATION window. No notice means no prep means no firewall rules on new host. Your ssh keys? Same? New host fingerprint? Verified? fail2ban? Installed? Firewall? Configured? What could go wrong: 1) old IP still responding, MitM 2) backup restored to shared storage, others peek 3) new hypervisor has host-level access to your memory 4) provider "oops" leaves your disk in old datacenter 5) their migration script logs your passwords by accident. Seen it all. Check your auth.log. Check it twice.

airgapped, encrypted, faraday'd, still worried
pablo

Member

#6

The same thing wey, pero with RackNerd, they send email at 3AM wey, migration at 6AM. I wake up 9AM and todo is broken. Their status page say "all systems green" pero my server is down wey? I check their ticket system, the timestamp say they knew 2 hours before wey. They lie to us? Que pasa con eso? Never again!

declarative or death
uma

Member

99.99% or bust

#7

Berlin1 said:
Their own ticket timestamps show restore completed six hours AFTER they declared "all clear."
Their status page showed 100% uptime for that day. My external monitoring (uptime robot + pingdom + self-hosted gatus) logged 14h 23m outage. I have the alert fatigue to prove it. 99.98% SLA? They calculate from their own status page, not reality. But providers do.

436 days. reboot is surrender.
cleardmitri

Moderator

Moderator

#8

Which backup frequency did you have configured?

No logs, no proof. I have logs.
PO

Member

#9

Their own status page doesn't count, surely

Post a reply

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

Post reply Preview Save draft