We ran into an interesting problem today while running bosh-init (2021 update: same error, but with bosh create-env) against AWS:
CPI ‘has_vm’ method responded with error: CmdError{“type”:”Unknown”,”message”:”AWS was not able to validate the provided access credentials”,”ok_to_retry”:false}`
This is a CPI error so after a bit of investigation confirmed our AWS keys were correct and valid with the awscli. After some more digging we found that the server we were running the bosh-init command from had drifted 7 minutes making the authentication token expired.
Update the ntp server configuration or manually move the minute hand on the server clock and get back to bosh-init glory.
You can use the following to sync of your time but best to fix your ntp settings:
sudo date -s "$(wget -qSO- --max-redirect=0 google.com 2>&1 | grep Date: | cut -d' ' -f5-8)Z"
date