[
https://issues.apache.org/jira/browse/WHIRR-88?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13051252#comment-13051252
]
Andrei Savu commented on WHIRR-88:
----------------------------------
Some of the things you could do improve the cluster startup time are:
* install java on the AMI and override install_java.sh with an empty function
* place the tarballs in /tmp/whirr/cache/files/ inside the AMI and upload them
to S3 to a container used for caching (whirr.blobstore-cache-container) and use
local paths in the .properties file. This will trick Whirr to believe that the
file have been uploaded and downloaded in a previous run / session and it will
use the ones that exists on the machine.
Probably it's not a bad idea to add a feature to Whirr so that it would be
possible to specify tarball URLs that are local to the remote machine (e.g.
remote:///data/zookeeper-3.3.3.tar.gz). It should be easy.
bq. Presumably, I should be able to ssh to a master created by Whirr and then
follow a cloud provider specific method for saving the current image.
Unfortunately it's not that easy now.
> Support image creation
> ----------------------
>
> Key: WHIRR-88
> URL: https://issues.apache.org/jira/browse/WHIRR-88
> Project: Whirr
> Issue Type: New Feature
> Components: core
> Reporter: Tom White
>
> Much of the time taken to start a cluster is in installing the software on
> the instances. By allowing users to build their own images it would make
> cluster launches faster. The way this could work is by having a create image
> step that brings up an instance and runs the install scripts on it before
> creating an image from it. The resulting image would then be used in
> subsequent launches.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
