-
Notifications
You must be signed in to change notification settings - Fork 169
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[Bug]: FAILED_TO_START_DUE_TO_NO_CAPACITY #1994
Comments
@judeleonard please provide One of the reasons can be that the fleet instance have GPUs while the dev environment doesn't request any. |
Here is the output
|
Yes, I tried to attach a GPU before but I got the same error 'Not having enough capacity' And my remote server actually has both GPU and docker preinstalled |
It's dstack by default offers only instances that match exactly the resources of the fleet |
This is the output. Not much details
|
Means the fleet creation wasn't successful.
This will help understand why fleet cound't be created |
I just recreated the fleet now
then my stack server log.Could this fingerprint be an issue with my ssh user?
|
Oh that's a known issue, it will be fixed in the next release but for now please do |
Thank you. Will do that |
hi @peterschmidt85 , sorry for updating you at this time. Our remote server was undergoing some updates. So I later tried it after installing paramiko like you suggested and dstack server logs indeed changed from what I had before. This is my log now. I also updated dstack to the latest v0.18.25. |
@judeleonard Now that you've updated |
dstack apply -f fleet.dstack.yml
dstack apply -f dev_environment.yml
|
But what it showed then? Was it successful? Running anything before the fleet is created doesn't make sense. Lets try to understand why the fleet isn't created. Need logs for that. |
|
This clearly shows that dstack cannot connect to the instance using the provided key @judeleonard Can you connect to the same host using the provided key via @un-def Any ideas what could be wrong? |
Yes, I can connect to the same server via ssh from my terminal. |
The current user I am using actually requires a password to successfully connect to the server. Could this be why? |
@judeleonard Yes! This certainly can be a reason. |
Okay, let me work on this and try it again. |
@judeleonard Also ensure the SSH key is added to |
Basically, |
Hello!
After applying with the command: "dstack apply .dstack.yaml" i have following:
Yes, minor bug, it completely ignored declared resource request with CPU 7, 10GB RAM and 15GB Volume, anyway not the main concern yet. Fleet is created, sudo is passwordless, docker is on the board: "Docker version 27.3.1, build ce12230".
So, seems doable. Then i ran one of examples i found in your docs section:
|
Hi @Fake45mar. Looks like your resources:
cpu: 2
memory: 2GB
disk: 5GB
gpu: 0 However, the instance in your FLEET INSTANCE BACKEND RESOURCES PRICE STATUS CREATED ERROR
fleet1 0 ssh (remote) 8xCPU, 39GB, 35.9GB (disk) $0.0 idle 1 min ago This instance does not match the Try removing resources:
cpu: 2..
memory: 2GB..
disk: 5GB..
gpu: 0 This notation means "2 CPUs or more", "2 GB or more", etc. You can find more examples of setting |
@jvstme , hey! Would like to thank you for the quick reply. Yes, it stopped failing with the code i mentioded before. However, it doesn't wake up vscode as it was supposed:
what it makes even more strange, is that :10999 port is not actually busy. Should i continue posting my questions here or it's better to open a new thread? |
@Fake45mar, yes, could you please open another issue for this question or join our Discord server? We can answer your question there. Happy to help! |
Steps to reproduce
create a fleet.dstack.yml to provision remote backend from my on_prem server. This was created successfully.
dstack apply -f fleet.dstack.yml
create another yml config to provision a development server with the provisioned fleet as the backend
dstack apply -f dev_environment.yml
Actual behaviour
I got the below error
All provisioning attempts failed. This is likely due to cloud providers not having enough capacity. Check CLI and server logs for more details.
Then also tried to see extra details about the error with the below command
dstack ps --verbose
output below
Expected behaviour
Instance provisioning should be completed successfully with a vscode link to my workspace.
dstack version
0.18.22
Server logs
Additional information
No response
The text was updated successfully, but these errors were encountered: