[
https://issues.apache.org/jira/browse/USERGRID-197?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14068634#comment-14068634
]
Yigit Sapli commented on USERGRID-197:
--------------------------------------
Unfortunately this is not true [~akarasulu].
Property substitution from ~/.m2/settings.xml only works locally. And these
credentials are apparently needed at cluster instances, which requires special
injection of these values, similar to the way cluster "_PRIVATE_HOSTS",
"_ADDRS" environment variables are injected in
webapp.coordinator.rest.CoordinatorUtils' executeClusterSSHCommands method.
However, I am noticing that passing of user sensitive information may not be
limited by AWS credentials. So, maybe instead of specifically injecting AWS
credentials, we should allow variable substitution in stack.json, reading from
local environment variables or from settings.xml. Then we can pass variables
like this in stack.json;
...
"scriptEnvironment" : {
"AWS_S3_ACCESS_KEY" : "$AWS_S3_ACCESS_KEY"
}
...
Dave, I can implement this after Wednesday, if that's ok.
> [CHOP] Provide mechanism for providing AWS creds to cluster instances
> ---------------------------------------------------------------------
>
> Key: USERGRID-197
> URL: https://issues.apache.org/jira/browse/USERGRID-197
> Project: Usergrid
> Issue Type: Story
> Components: Chop
> Reporter: David Johnson
>
> In a cluster setup script, we need to be able to download things from S3 and
> so we need AWS credentials to be passed.
> We cannot specify them in the stack.json file because that file is to be kept
> in source code control.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
