Skip to content
Register

YABS on 50 different $10/year plans, the variance is absurd

VPS Hosting by petra 29 replies 4.8K views
lucgone

Member

Le Baguette

#11

petra said:
Some NVMe plans score worse than decade-old SSD

Contabo Nuremberg, mon dieu. J'ai eu un VPS là, le YABS disque était 400 MB/s. NVMe. Quatre-cents. Mon SSD SATA vieux de huit ans fait mieux dans le garage.

Vive la résistance... électrique
mediaaustin

Member

Deliverability Nerd

#12

Following this thread closely. Disk variance matters for mail queues too. A slow commit on postfix means deferred mail piles up. Not all $10/year plans can handle even modest volume.

SPF, DKIM, DMARC — holy trinity ✉️
FL

Member

#13

Same here on the Contabo lottery. Got a St. Louis instance, disk was fine, CPU was E5 v3. Friend got Seattle, same plan, Ryzen 5 3600. Night and day.

marcus_qc

Member

#14

RAJ said:
Hostinger? Their Mumbai node was all over the place

Not my company so I won't speak for them, but we see this industry-wide. Plan name stays the same, hardware refreshes in waves. The resellers don't always know which wave they're on. Ask for node ID before you buy, some will tell you.

42U and still growing
NUT

Member

Bangkok

#15

Bangkok to Singapore is 35ms. GreenCloudVPS Singapore sometimes good, sometimes not. I stay with Vultr Singapore now, more expensive but YABS is stable month to month. Four servers, cannot afford lottery.

paolo62

Member

#16

Ramen-funded so I buy annual. RackNerd LA was great in 2022, renewed 2023, new node, disk halved. No refund on annual. Lesson learned.

MA

Member

#17

paolo62 said:
No refund on annual

This. The annual discount is a trap if hardware rotates underneath you. I only do monthly now even if it costs 40% more.

LichunLars

Member

#18

torontoeng said:
Falkenstein vs Helsinki

Stockholm to Falkenstein is 23ms, to Helsinki 17ms. But my Hetzner box is in Falkenstein and disk is consistently 1.2 GB/s. Friend in Helsinki sees 800 MB/s sometimes, 1.5 sometimes. The older generation you found, petra, explains it.

1ms or I don't want it
petra

Member OP

#19

LichunLars said:
Explains it

Exactly. In the dataset, Helsinki has two clusters: 2019 hardware at ~850 MB/s and 2022 at ~1.6 GB/s. Falkenstein is all one cluster, 2021-ish, around 1.2. So your experience tracks.

JA

Member

#20

Can someone explain why CloudCone only has Los Angeles? Single point of failure much?

Post a reply

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

Post reply Preview Save draft