HN Explorer

Ask HN: Terraform is getting a FOSS fork, what about other HashiCorp tools?

12 points by jimkoen ↗ HN
The recent license change from MPL to a non-FOSS license was discussed to great extend on HN over the past week [0]. As a result, some companies with infrastructure/devops focused products announced that they plan to creating a FOSS fork of Terraform, as it is arguably the most popular tool in HashiCorps toolshed. [1]

But I was wondering if anyone was interested in or is planning to fork some of the remaining tools in the HC stack. I'm personally interested in keeping Packer going, as it currently seems to be the only solution that can create what HC calls "golden images" in an user friendly way. An alternative to Packer would be shoehorning something with Ansible, which I think would be rather tedious.

I was hoping that there maybe was some ongoing development already, but of which the greater HN community wasn't aware of, so maybe this post could be used to spread some awareness.

[0] https://news.ycombinator.com/item?id=37081306 [1] https://news.ycombinator.com/item?id=37133054

3 comments

[ 3.3 ms ] story [ 14.6 ms ] thread
remram
Images created by Packer are not affected by the change of license, and I don't know if many products are built on the Packer code. Re-doing the work in a separate form just to have it open source might not be very worthwhile.

There's no need of a fork to "keep Packer going" as you put it. HashiCorp is continuing development, under a different license, but one that probably still works for everyone.

thayne
IANAL but that isn't obvious to me. What if you have an automated process to build images that you use in a hosted service product thay competes with a hashicorp product. I could see an argument that you are "embedding" packer in your service.

And any CI/CD product can no longer include building images with packer (at least as long as it includes a distribution of packer), or probably even have packer pre-installed.

remram
> What if you have an automated process to build images that you use in a hosted service product thay competes with a hashicorp product.

Yes agreed, those people would want to maintain a fork from the latest MPL release.

However it is not clear that those people exist. In addition, it is not clear that there are enough of those people, with different enough commercial product, such that they would benefit from collaborating in an open-source manner.

My point is, OP is probably not one of them.