Skip to content
This repository has been archived by the owner on Nov 23, 2017. It is now read-only.

Add README note about project status and recommended alternative #117

Merged
merged 2 commits into from
Nov 4, 2017

Conversation

nchammas
Copy link
Contributor

spark-ec2 is no longer under active development. This has been the reality for about a year or so, but I have been asked by @shivaram to submit a PR making this official by adding a note to the README and redirecting people to Flintrock, a project I created as a sort of spiritual successor to spark-ec2.

I think @shivaram can provide a proper post-mortem on his end for why spark-ec2 development has effectively stalled. From my perspective, I believe the main reasons are:

  • spark-ec2's dependence on custom AMIs makes it difficult for the project to keep up pace with Spark itself. This is a problem I tried unsuccessfully to tackle for spark-ec2 years ago and have discussed here as well.
  • Major improvements to spark-ec2's user experience (launch speed, persistent configs) require a lot of review time, and the core people on this project (mainly @shivaram these days, myself and others to a lesser degree) don't have the bandwidth to keep up.

Flintrock addresses these key issues -- they were, in fact, the main motivation for the project -- and I'd be happy to help onboard users who decide to switch from spark-ec2 to Flintrock.

@shivaram
Copy link
Contributor

+1 - Thanks @nchammas for the PR. I think the need for custom AMIs and tcode changes for every Spark release / EC2 instance release results in significant burden in terms of maintenance for spark-ec2. I think Flintrock's architecture avoids some of these issues and there are also more mature commercial tools like EMR Spark or Databricks for more advanced use cases.

I'll leave this PR open for a couple of days to see if there are any other comments.

@nchammas
Copy link
Contributor Author

nchammas commented Nov 2, 2017

I should probably push additional commits to this branch to add PR and Issue GitHub templates so that when people file new issues or open new PRs they are told that this project is inactive and are redirected to the README.

@shivaram
Copy link
Contributor

shivaram commented Nov 3, 2017

That would be great @nchammas -- Thanks

@shivaram
Copy link
Contributor

shivaram commented Nov 4, 2017

LGTM. Thanks -- Merging this.

@shivaram shivaram merged commit f4bc534 into amplab:branch-1.6 Nov 4, 2017
@nchammas nchammas deleted the flintrock-pointer branch November 4, 2017 18:32
@nchammas
Copy link
Contributor Author

nchammas commented Nov 8, 2017

@shivaram - Looks like GitHub recently added the ability to archive repositories. That will obviously work much better than the PR and Issue templates I added here.

Perhaps it's a good idea to make a final sweep through any open issues and PRs to leave comments where appropriate and then archive this repo.

@shivaram
Copy link
Contributor

shivaram commented Nov 8, 2017

Thanks for the pointer. I will take a look at it.

@shivaram
Copy link
Contributor

@nchammas Given that the oldest PRs are more than 6 months old I'm not sure its useful to go and ping them again. I think archiving just makes the PRs and issues read-only which should be fine ?

@nchammas
Copy link
Contributor Author

@shivaram - That's fair. I was thinking of the more recent issues and PRs where the authors might still be expecting a response, but I suppose they will find out sooner or later when they check and see that the repo was archived.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants