Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Mitchell reflects as he departs HashiCorp

hashicorp.com

Post reply on HN
Print view View original

81–90 of 192 posts

Re: Mitchell reflects as he departs HashiCorp

#81

Post by acdha »

nuker wrote: post #71

Just a side thought, with all due respect: > The controversial worldviews such as multi-cloud that we founded this company on are now mainstream and broadly accepted. Emm .. no. It is not mainstream. Multi-cloud may be forced choice for some regulated fintech, but no sane project will double the infra codebase just to get from "lock-in". I see no other benefits. And double the codebase is quadruple the bugs. This is…

So I kind of agree that most projects don’t need multi cloud but any enterprise will have a bunch of different things which can be managed by Terraform. There is real value in being able to use the same tool to manage AWS/GCP/Azure/etc., Cloudflare, GitLab, VMware, and even racks of Cisco kit in the basement.

If you’re doing it that way, you avoid most of those multicloud drawbacks because you’re using the full native functionality, not building abstraction layers or racking up technical debt by sticking with the lowest common denominator.

Re: Mitchell reflects as he departs HashiCorp

#82

Post by leriksen »

nuker wrote: post #71

Just a side thought, with all due respect: > The controversial worldviews such as multi-cloud that we founded this company on are now mainstream and broadly accepted. Emm .. no. It is not mainstream. Multi-cloud may be forced choice for some regulated fintech, but no sane project will double the infra codebase just to get from "lock-in". I see no other benefits. And double the codebase is quadruple the bugs. This is…

That's not what he means when he says multi-cloud.

He means his product supports multiple clouds, through terraforms providers, for example.

Doing the same infra, across multiple clouds, is a different thing, it's not a Hashi thing.

Re: Mitchell reflects as he departs HashiCorp

#85

Post by aeyes »

fangorn wrote: post #62
echelon wrote: post #5

Earlier quoted context omitted.

They built a ton of value, were unable to capitalize on it (just like early Docker), and when they tried to capture the value of the thing they built it pissed off the open source ecosystem (and all the profitable companies) built on top. They should have thought about this a long, long time ago. I feel for the smaller companies, but I feel for big companies that come in and plunder because of open license terms.

I recently concocted a (conspiracy) theory that relicensing of HC projects is a ploy to get IBM (or some other company happy to rain on IBM's parade) to finally make an offer. Mitchell leaving the company features in the theory as well. Basically it seems like since at least 2021 HC leadership and/or investors are exploring exit strategies that will bring in the beeeelions.

Hashicorp is a public company so the theory doesn't make much sense to me. The execs and early investors already made their money.

Re: Mitchell reflects as he departs HashiCorp

#86

Post by nuker »

acdha wrote: post #81
nuker wrote: post #71

Just a side thought, with all due respect: > The controversial worldviews such as multi-cloud that we founded this company on are now mainstream and broadly accepted. Emm .. no. It is not mainstream. Multi-cloud may be forced choice for some regulated fintech, but no sane project will double the infra codebase just to get from "lock-in". I see no other benefits. And double the codebase is quadruple the bugs. This is…

So I kind of agree that most projects don’t need multi cloud but any enterprise will have a bunch of different things which can be managed by Terraform. There is real value in being able to use the same tool to manage AWS/GCP/Azure/etc., Cloudflare, GitLab, VMware, and even racks of Cisco kit in the basement. If you’re doing it that way, you avoid most of those multicloud drawbacks because you’re using the full nativ…

> There is real value in being able to use the same tool to manage AWS/GCP/Azure/etc., Cloudflare, GitLab, VMware

My bad, I did not define meaning of multi-cloud. I meant using AWS and GCP at the same time, doing mostly the same things, for redundancy and no lock-in.

Still I prefer native IaaC tools, it is simpler, faster and more reliable. I'm AWS guy, so Cloudformation for all things infra. It irks me when I see Terraform doing AWS infra. I jump in to rewrite it as first matter of business, lol

Re: Mitchell reflects as he departs HashiCorp

#87

Post by candiddevmike »

fangorn wrote: post #62
echelon wrote: post #5

Earlier quoted context omitted.

They built a ton of value, were unable to capitalize on it (just like early Docker), and when they tried to capture the value of the thing they built it pissed off the open source ecosystem (and all the profitable companies) built on top. They should have thought about this a long, long time ago. I feel for the smaller companies, but I feel for big companies that come in and plunder because of open license terms.

I recently concocted a (conspiracy) theory that relicensing of HC projects is a ploy to get IBM (or some other company happy to rain on IBM's parade) to finally make an offer. Mitchell leaving the company features in the theory as well. Basically it seems like since at least 2021 HC leadership and/or investors are exploring exit strategies that will bring in the beeeelions.

They've been for sale since pre-IPO. Cisco almost bought them.

Re: Mitchell reflects as he departs HashiCorp

#88

Post by nuker »

leriksen wrote: post #82
nuker wrote: post #71

Just a side thought, with all due respect: > The controversial worldviews such as multi-cloud that we founded this company on are now mainstream and broadly accepted. Emm .. no. It is not mainstream. Multi-cloud may be forced choice for some regulated fintech, but no sane project will double the infra codebase just to get from "lock-in". I see no other benefits. And double the codebase is quadruple the bugs. This is…

That's not what he means when he says multi-cloud. He means his product supports multiple clouds, through terraforms providers, for example. Doing the same infra, across multiple clouds, is a different thing, it's not a Hashi thing.

> Doing the same infra, across multiple clouds, is a different thing, it's not a Hashi thing.

Not sure about this. Here is what official site says:

"Provisioning infrastructure across multiple clouds increases fault tolerance, allowing for more graceful recovery from cloud provider outages."

https://www.terraform.io/use-cases/multi-cloud-deployment

Re: Mitchell reflects as he departs HashiCorp

#89

Post by acdha »

nuker wrote: post #86
acdha wrote: post #81

Earlier quoted context omitted.

So I kind of agree that most projects don’t need multi cloud but any enterprise will have a bunch of different things which can be managed by Terraform. There is real value in being able to use the same tool to manage AWS/GCP/Azure/etc., Cloudflare, GitLab, VMware, and even racks of Cisco kit in the basement. If you’re doing it that way, you avoid most of those multicloud drawbacks because you’re using the full nativ…

> There is real value in being able to use the same tool to manage AWS/GCP/Azure/etc., Cloudflare, GitLab, VMware My bad, I did not define meaning of multi-cloud. I meant using AWS and GCP at the same time, doing mostly the same things, for redundancy and no lock-in. Still I prefer native IaaC tools, it is simpler, faster and more reliable. I'm AWS guy, so Cloudformation for all things infra. It irks me when I see Te…

Heh, I have mostly worked on AWS but we stopped using CloudFormation because the experience was so much worse than Terraform with the lengthy deploys, no diffs, significant time delays supporting new AWS features, and deadlocks. They’ve added diffs but turning debugging into a half hour or longer break is still a problem.

Re: Mitchell reflects as he departs HashiCorp

#90

Post by scoot »

> the GitHub Octoverse report found that HashiCorp Configuration Language (HCL) has once again emerged as one of the top languages used in open source projects

"Hashicorp Configuration Language (HCL) with a 56,1% increase in popularity is the fastest-growing language according to GitHub"

Fastest growing and "top" are not the same thing, but I have to assume that you know that @mitchellh. Embarrassed for you.

https://xkcd.com/1102/

Post reply on HN