New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Scaleway Stardust IPv6 €0.37/mo
So I created the Stardust instance on Scaleway. It was out of stock but the API way worked fine.
Performance is very good for the price. Highly recommended.
YABS:
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
# Yet-Another-Bench-Script #
# v2022-02-18 #
# https://github.com/masonr/yet-another-bench-script #
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
Tue Apr 5 14:17:44 UTC 2022
Basic System Information:
---------------------------------
Processor : AMD EPYC 7281 16-Core Processor
CPU cores : 1 @ 2096.062 MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ✔ Enabled
RAM : 973.3 MiB
Swap : 0.0 KiB
Disk : 8.9 GiB
fio Disk Speed Tests (Mixed R/W 50/50):
---------------------------------
Block Size | 4k (IOPS) | 64k (IOPS)
------ | --- ---- | ---- ----
Read | 150.23 MB/s (37.5k) | 816.09 MB/s (12.7k)
Write | 150.62 MB/s (37.6k) | 820.38 MB/s (12.8k)
Total | 300.86 MB/s (75.2k) | 1.63 GB/s (25.5k)
| |
Block Size | 512k (IOPS) | 1m (IOPS)
------ | --- ---- | ---- ----
Read | 1.14 GB/s (2.2k) | 1.12 GB/s (1.0k)
Write | 1.20 GB/s (2.3k) | 1.19 GB/s (1.1k)
Total | 2.35 GB/s (4.5k) | 2.32 GB/s (2.2k)
iperf3 Network Speed Tests (IPv6):
---------------------------------
Provider | Location (Link) | Send Speed | Recv Speed
| | |
Clouvider | London, UK (10G) | 359 Mbits/sec | 543 Mbits/sec
Online.net | Paris, FR (10G) | 211 Mbits/sec | busy
WorldStream | The Netherlands (10G) | busy | busy
WebHorizon | Singapore (400M) | 336 Mbits/sec | 413 Mbits/sec
Clouvider | NYC, NY, US (10G) | 256 Mbits/sec | 1.86 Gbits/sec
Clouvider | Los Angeles, CA, US (10G) | 239 Mbits/sec | 1.08 Gbits/sec
Geekbench releases can only be downloaded over IPv4. FTP the Geekbench files and run manually.
Comments
That's really impressive for the price.
Can you show me pricing page for it ?
Order the regular Stardust for 0.0025€/h (1,8€/pm) and remove the IPv4. Done.
Yep, that's it.
Command to create the instance using API (read the docs):
On their site they mention that these instances are very limited. Is that just marketing, or can you expect that API call to work only infrequently?
I think it works most of the time, however, I think old accounts are limited to two instances and new accounts one instance.
My API call fails, but I think that's because I already have one instance, and I'm not in the mood to destroy it in order to test my theory
Edit: Speculating a lot since I really don't know. I just know the API worked for the instantly but once I had one it just didn't (many people say that it works until you've got one in par-1 and one is ams-1 but for me it dosen't work and I've only got 1 in ams-1 and 0 in par-1).
It was supposed to be one per DC: AMS and PAR. Try creating in another DC.
Edited the post. It dosen't.
Error:
Bad image label 'ubuntu_focal' for STARDUST1-S
I have created instance but what will be the default root password?
You'll have to upload an SSH key to Scaleway. If you don't have native IPv6, use Cloudflare Warp when connecting
What location? AMS?
i have just run your script and it worked may be netherland.
Well, could be this OS is not available in the particular DC? Weird, but maybe just try with Debian? Tnx
Bad image label 'debian_bullseye' for STARDUST1-S
All I can say it is a weird error to say the instances are out of stock, or that your quota is out.
Well everything from the error code to their marketing and them not mentioning the limits all seems weird to me.
Ohh dm
I have 2 (one in each DC) but it has been idling for as long as I remember.... any good uses for it?
Any web-related thing. Cloudflare does the magic and will allow IPv4 & IPv6 to connect.
I use mine for monitoring response times and DNS queries.
Hey, thanks for the tip and trying this myself. is it enough to turn off the public ip in the Scaleway console or is anything more needed to "remove the IP4"?
If you use the API command (scw instance server create type=STARDUST1-S zone=nl-ams-1 image=ubuntu_focal root-volume=l:10G name=test ip=none ipv6=true project-id=YOUR-PROJECT-ID) it creates it without IPv4, see "ip=none ipv6=true".
Scaleway just came up with a current consumption pie of €0.04 excl. taxes, €0.01 for compute and €0.03 for Local SSD. This is the basic 10GB local storage created initially. This comes to €1.2 pm (assuming no further charges in the balance day) and about €15 per year. Not cheap for an ipv6 only vps.
The current consumption is just an estimate and does not scale in the way you are calculating. It may stay on those values for several days.
Since I cannot select a Stardust instance to create and look at the estimate right now, I will refer to my bill and do some calculations manually.
Pulling out last months bill, my compute charge for a single Stardust instance was €0.09:
A local SSD is billed at €0.0004/hour for 10GB. This comes out at €0.29.
You can find this by trying to create one from the dashboard (Make sure not to select Block storage!)
29 + 9 = 38 so the monthly cost of a Stardust without IP is €0.38/mo. This is consistent with what I thought it was.
Apply that annually and you get €4.56/yr
The Stardust instances are weird because they're wildly cheaper than the other instances at Scaleway.
Digging into it further to find the hourly for the compute alone:
The website advertises the total cost as €0.0025/hour
We can know from the console that an IP adddress is €0.002/hour so that already brings our total down to €0.0005/hour
We know from earlier the SSD is €0.0004/hour
Therefore it stands to reason that the cost of the compute is €0.0001/hour
I hope that clears things up, I'm not really all that good at crunching numbers like this so thats a bit all over the place, but it's all consistent with what I've experienced over the last year of running a pair of Stardust instances without IP addresses.
Great calculations, but, I also think it's unfair to compare this to a IPv6-only one-man-show operation and say that this is a dollar more expensive and therefore 30 % more expensive, Scaleway, unlike the IPv6-only one-man-shows, isn't going away tomorrow.
As long as I'd have a decent backup system and decent monitoring, I'd feel great running production (not hyper important) stuff on the Stardust. Especially if you run both snapshots as well as off-site backups.
The most probable bad thing that could happen would be a disk failure, well, a boot from a snapshot takes a minute.
Sucks though that they don't do automatic snapshots, but hey, there's solutions (3rd parties & API) for that too.
Thank you guys for your insights. I only have a day's experience and was extrapolating but true, it may not scale like that. Stardust is by no means my production server atm but looking ahead, costs permitting, could definitely make more use of it.
for me it creates it with api but then refuses to start it.
wait nvm I used a different region this time and it worked!
Thank you, this worked using the CLI
also you can change the image to a different OS in cli:
Instead it's going to burn down tomorrow