Using the docker alpine 2.0 ce version, I have implemented a mechanic where certain configuration information is retrieved from aws dynamodb when a new container instantiates (auto-scaling).
Using pl.utils.execute() we can successfully call the aws cli command and retrieve a response (having passed the credential information into the environment variable -e
“AWS_SHARED_CREDENTIALS_FILE=/usr/local/credentials” during run) when called during the “new” phase.
However, when we attempt these calls during the “access” phase (understanding the performance implication), I get the "Unable to locate credentials. You can configure credentials by running “aws configure” error.
Disregarding whether it is wise to make a blocking http request durign the access phase, any suggestion/ideas of why the “aws cli” call would fail during :new as composed to :access? My guess is that its related to the master vs worker processes in nginx.
Thank you in advance for any help.