Skip to content
Register

RackNerd: mixed bag, fast and cheap but occasional outages

Reviews by petra 7 replies 453 views
petra

Member OP

#1

We've been with RackNerd for 18 months now. Wanted to share an honest take since these forums tend toward extremes.

The good: NVMe is actually fast. Consistent 450+ MB/s on our workload. Price is industry-standard for the specs. Network is solid in EU.

The bad: had three unplanned outages. 12 min, 47 min, and 8 min. No root cause published. Support acknowledged "upstream provider maintenance" each time. Uptime calculator says 99.7% but feels worse when it clusters.

The weird: their control panel randomly logs you out. No pattern. Just gone.

Would I recommend? With caveats. Not for production without failover. :)

tallinnying

Member

Night Shift

#2
petra said:
Not for production without failover.

Wat man is not fast is more fast wen is working but wen is gone is gone man kkkkkk

I no have failover so is pain for me

Not gonna say is bad not gonna say is good man

builds at 3AM, sleeps at noon
engzur

Member

BGP Dreamer

#3

AS6939 checking in. Peered with them at three exchanges. Their prefix stability is mediocre at best. Saw a 15-minute flap at AMS-IX last quarter. No IRR entry for the maintenance window. If you don't have your own ASN and at least a /24 you're not even in the game.

[RackNerd]===flap===>[AMS-IX]===drop===>[you]
| |
No RPKI no route

Get real transit or accept the risk.

pieter_rtm

Member

#4

RackNerd specs we measured:

  • CPU: Geekbench 5 single 890, multi 3240
  • Disk: 480 MB/s sustained, 220 MB/s after 30s write
  • Network: 2.3 Gbps burst, 850 Mbps sustained
  • Latency AMS-SIN: 168ms, AMS-SAO: 198ms

Three outages in 18 months is 99.92% not 99.7%. Their math is wrong. As said, control panel session timeout is 4 hours but their JS timer drifts. Known issue, no fix scheduled.

Dry: you get what you pay for. It is cheap.

Containers before it was cool
jane_ffm

Member

#5

Thanks for the balanced review, @petra. I appreciate honest feedback in this space.

A few notes from my side — Contabo publishes root cause analyses within 72 hours for any incident over 5 minutes. I'd love to have you take a look at their uptime commitments.

Not here to poach, just to say that session timeout reliability and transparent post-mortems are table stakes for them. Happy to check whether their Nuremberg or Düsseldorf locations would improve those latency numbers for your use case — feel free to DM.

— Jane, ex-support

Single mode till I die 💀
RAJ

Member

#6

What workload you running to hit 450+ MB/s sustained

your margin is my opportunity
SU

Member

#7

I had the 47 min one too, no explanation given

minh1987

Member

#8

99.7%? Three outages and no root cause, nah

phở at 3AM, deploy at 4

Post a reply

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

Post reply Preview Save draft