All tenancies get six public IPv4 addresses for Always Free 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.
But so far, I’ve only managed to get 3 (2 ephemerials, 1 reserved).
Maybe I’m not understanding this correctly. Is there any resource online that further explains this part?
Seems like they changed this, now the page doesn’t say anymore “All tenancies get six public IPv4 addresses for Always Free compute instances.”. It doesn’t specify the number anymore.
All tenancies get two public IPv4 addresses for Always Free compute instances. If you want to create more than two Always Free instances, you can create the instances without assigning public IP addresses.
Also, if you just want access to the VM, you should really try out their ‘Bastion’ service. It is a great tool to get secured access to VMs in private subnet (even other ports through port forwarding)
As for the errors I was getting. I did get the error message saying I am only allowed 2 ephemerial public IPs and 1 reserved public IP.
So I think those were pretty clear limitations. I was just confused by the Always Free Resources documentation that I listed because it says there the we can have 6 public IPs.
Yeah but I think the ephimereal ips are not counted. Because that changes on restart. And I still only have 1 static public IP. But I gave up on this a long time ago. Thanks for the response tho.
What are you supposed to mean with this image? I can see a mention to “bring your own IP” and pools of IP addresses, and none of those are related to your question.
Oh, I just skimmed the VCN section. Okay after setting scope to my region I did see the reserved IP limitation. And yes, always free tier only get 1 reserved public IP. As for the ephemerial, even though it wasn’t listed in the Limits page, it’s pretty clear that I only get 2 based on the error that I get.
So that quote from the documentation really confuses me.
How are you filling the 6 IP addresses, btw? We get only 2 instances as always free, isn’t it?
We actually get up to 4 ARM instances. The resource limit is 4 cores and 24 gigs of RAM. So you can spin up 4 vms with 1 core each and just divide the ram allocation as needed.
I can only get ARM instances. I was never lucky enough to get the AMD instances.
No, it’s possible to create max 4 free instances.
And there’s load balancer which needs one IP as well I think. I only used 2 ephemeral & 1 reserved IP in my 2x AMD & 1 ARM instance though.
Just on the additional static IPs. I used to run multiple instances so I wanted that. But now that I settled for one or two and situational, additional static IPs don’t matter anymore.
Yes, the doc is confusing if you get error after 3 IP addresses. I hope someone else will pitch in who managed to run into the same problem and fix it (?)
Interesting usage there with ARM instances.
Like I mentioned in my parallel comment, you really don’t have to allocate public IP address to all. If you just want access into the machine, you can use the bastion service (they’re time limited, but free).
Maybe a workaround for non production environments would be to write a script that runs on VM startup, checks if the ephemeral IP has changed, and if so, sends a request to the DNS/registrar to update that DNS record.
That would mean some downtime after reboot (waiting for the DNS propagation) but could be acceptable for staging/testing environments.
Oh, it’s just that I was hoping to host web services on the machines so I needed the public IPs.
But 3 is still a lot for a free service. I suppose I can just reallocate the 4th VMs resources to the other VMs.
The ARM VMs are actually pretty good at least for my use case. I still hope to snag some AMD instances so I still check from time to time. But I never really saw one since I registered in Oracle Cloud. So maybe it’s just not available in my region.