skypilot-users

How can I solve the 'wait_ready timeout exceeded' error when running sky spot launch?

ol

oliver jamie

Asked on Sep 21, 2023

This error usually occurs due to a networking issue. Here are a few steps you can take to troubleshoot:

  1. Check network connectivity to any AWS IP and ensure that VPN is not required.
  2. Verify if any networking or VPC settings are set in ~/.sky/config.yaml.
  3. Try running 'sky launch' instead of 'sky spot launch' to see if it works.

If the issue persists, you can try the following:

  1. Check if you can SSH into the IP directly using the command 'ssh ubuntu@<IP_ADDR> -i ~/.ssh/sky-key'.
  2. Update SkyPilot to the latest version (0.4 or skypilot-nightly) and try launching the cluster again.

If none of these steps resolve the issue, you can try launching an AWS instance manually from the console and see if you can SSH into it. If that works, it could indicate a firewall issue on your local machine.

Please note that the key should be added directly to the authorized keys in the VM by SkyPilot, rather than using a key pair.

If you are still facing issues, it is recommended to reach out to the SkyPilot support team for further assistance.

Dec 19, 2023Edited by