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.
Comments
Yep, his were among the main ones I watched when getting my Oracle setup going.
Haven’t really messed with my OCI free tier VMs in a few years but pretty sure my original non-ARM vms are still humming along.
A few questions:
Yes.
200 GB total across all VMs, including your old non-ARM ones.
No.
Public IPs are still available (for now) and the amount you can have was raised slightly when they added ARM instances (Though I don't remember the exact numbers)
Yes. It is still 10 TB.
Exactly, for ARM its not possible, but for x86 its random - some nodes have it enabled. I have two x86 instances, I have nested virt on one. If someone needs this flag then try to provision instance once again if you dont have it
Very true, thank you! Indeed whether you get nested virt or not on Epyc machines is essentially random.
Rather than provision a new instance to reset your odds, you should be able to stop and start it again, since it may start on a new node (The disk is block storage and not local). There are cases where people have nested virt and then stop/start and don't have it anymore, the same can be true the other way.
Can we use it all that way in non ARM VMs ?
The answer sadly no, you can only have 1 core and 1GB ram for one single VM.
Ok I figured
Has anyone ever used an OCI ARM instance for either of the below?
Just wondering if anyone has successfully setup either of these scenarios as I know their could be challenges around ARM in potentially setting these up. If anyone knows of any online guides/tutorials/tips they have utilized that would also be awesome.
A cursory Google search doesn’t reveal too much as far as guides for ARM for Nextcloud but admittedly I haven’t done any thorough digging yet. If I find any guide that works I will also post it.
Thanks!
For most things it makes not difference that linux is running on ARM instead on x86 arch, in the worst case, you might need to compile an application yourself. I did setup a nextcloud instance for testing and i didn't had any issues.
If you want a tutorial, add 'raspberry pi' to your search term - if something can be done on Raspberry Pi, it should be possible to do on a oracle cloud ARM instance.
I've installed Hestia on the ARM versions and it ran smoothly.
Because for both of these tasks there doesn't need to be any guide "for ARM", just go with any generic one.
So I logged in to my account after having neglected the account and my VMs for a few years, and it says to use any free tier resources, I need to upgrade to a paid account. Is this normal? How does that make any sense, do they just want you to enter a CC I guess?
What's the exact phrasing? I haven't seen that
It’s a banner at top:
“You are using a Free Tier Account. To access all services and resources, upgrade to a paid account.”
So I was being dumb I think this normal— I forgot I had to choose my tenant name to see or create anything… looks like I can create new instances within free tier limits..
In case anyone is interested, just spun up a 4 OCPU 24G Ampere instance in Ashburn A2:
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
Yet-Another-Bench-Script
v2022-06-11
https://github.com/masonr/yet-another-bench-script
## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##
Sun Jun 26 23:32:56 UTC 2022
ARM compatibility is considered experimental
Basic System Information:
Uptime : 0 days, 0 hours, 42 minutes
Processor : Neoverse-N1
CPU cores : 4 @ ??? MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ❌ Disabled
RAM : 23.4 GiB
Swap : 0.0 KiB
Disk : 96.8 GiB
Distro : Ubuntu 22.04 LTS
Kernel : 5.15.0-1011-oracle
fio Disk Speed Tests (Mixed R/W 50/50):
iperf3 Network Speed Tests (IPv4):
Provider | Location (Link) | Send Speed | Recv Speed
| | |
Clouvider | London, UK (10G) | 872 Mbits/sec | 768 Mbits/sec
Online.net | Paris, FR (10G) | 1.14 Gbits/sec | 1.65 Gbits/sec
Hybula | The Netherlands (40G) | 983 Mbits/sec | 2.07 Gbits/sec
Uztelecom | Tashkent, UZ (10G) | busy | 1.07 Mbits/sec
Clouvider | NYC, NY, US (10G) | 3.40 Gbits/sec | 4.01 Gbits/sec
Clouvider | Dallas, TX, US (10G) | 578 Mbits/sec | 829 Mbits/sec
Clouvider | Los Angeles, CA, US (10G) | 1.04 Gbits/sec | 1.14 Gbits/sec
Geekbench 5 Benchmark Test:
Test | Value
|
Single Core | 853
Multi Core | 3336
Full Test | https://browser.geekbench.com/v5/cpu/15685031
Create x86 instance please, Im interested which regions get 7742 or 7J13... Frankfurt and Amsterdam only 7551.
I saw 7742 is possible in Tokyo.
I have two in London, both are 7551. Doesn't prove much though.
There should be 7742 and 7J13 according to Oracle docs, but of course they don't specify in which regions xD I only know about 7742 in Tokyo, because I saw yabs on some japanese blog. Never saw 7J13, but it should be possible...
"VM.Standard.E2.1.Micro: E2-based, E3-based, or E4-based standard compute. Oracle Cloud Infrastructure assigns one of the following processors:
AMD EPYC 7551. Base frequency 2.0 GHz, max boost frequency 3.0 GHz.
AMD EPYC 7742. Base frequency 2.25 GHz, max boost frequency 3.4 GHz.
AMD EPYC 7J13. Base frequency 2.55 GHz, max boost frequency 3.5 GHz."
https://docs.oracle.com/en-us/iaas/Content/Compute/References/computeshapes.htm
Looks like it's not allowing any AMD instances to be spun up in Ashburn AD1/2/3 (at least for me)
Out of capacity?
But what's the point, the 1/8 of core that you get would be slower than the ARM instance in any case.
Used to be 4 IPv4 addresses per instance, not anymore as they have changed to:
IP Addresses for Compute Instances
You do not have to assign a public IPv4 address to every compute instance in your tenancy. You can create a compute instance in a public subnet without assigning the instance a public IP addresses, and create an instance in a private subnet.
if you have added previously under old terms and attached to instance they unable to remove them.
What you quoted doesn't say you can't assign 4, just that you don't need to assign a public IP, same as always.
However, I believe the free tier resources page was updated and doesn't mention the reserved and ephemeral IP's limit like I last recall seeing.
Used to state the limitation of public IP address per instance, instead of my quoted text.
You still get a public IPv4 address but you can just enable IPv6. You get a /54 which you can then allocate to your instances.
Has anyone actually been capped?
It says 10TB egress but from my tests there is no limit yet. Not in Amsterdam or Johannesburg.
it's possible to deploy VMs across multiple locations?
Yeah, sure. You have to upgrade your account to pay as you go account. Always free resources are limited to home region only.
When you upgrade to pay-as-you-go, any other thing one needs to be aware of? Would like to avoid surprises
Just don't use/provision anything without knowing about it & it's pricing.
After all it's cloud just like AWS or GCP. If you don't use any paid resources it's better not to upgrade PYG account.