What are potential alternatives for Mist.io?

[olljanat]

Before going too far with this project, I want to also make sure that there is no better alternative project contribute to.

So would be good if someone can help to understand which other open source tools there are why not to contribute to them instead of?
Especially tools coming from Asia are interesting because they have a lot of development resources and many of those tools are just missing proper translation and documentation in English.

So far I have seen these but not tested yet:

• https://cloudforet.io (This seems to be even part of Linux Foundation)
• https://www.cloudpods.org (build by Chinese)

[olljanat]

@philcryer @Ben0p unfortunately it starting to looks that getting Mist working properly is very big task.

Alternatives investigate so far:

• Remove most of the feature so that features can be same for all Cloud providers: Remove broken Cloud providers and incompatible networking features #22
• Going back to Mist heavily modified version of libcloud Revert libcloud back to Mist.io customized version and remove broken Cloud providers #23 and agree that it will be fully incompatible it upstream.

Some point I actually tested Cloudpods earlier and contributed a bit to their code yunionio/ocboot#1418 (kind of test that how they handle pull requests).

For some reason I didn't liked it too much but probably will go test it again and potentially test Cloudforet too.

Feel free to try those version is pull requests and comment which way forward is better for this project in case it continues.

[olljanat]

https://www.cloudpods.org/ (build by Chinese)

Quickest way to setup Cloudpods is to follow this guide.

Just run these commands first so you will download images from Docker Hub instead of Chinese servers:

export CLOUDPODS_REPO=docker.io/yunion
export MARIADB_REPO=docker.io/library

These clouds seems to be supported out of the box:

To GCP I was not able to connect and with Azure VM create feature is not available but with AWS, Alibaba Cloud and ProxmoxVE creating VMs works fine.

Biggest challenges are that UI is quite complex as there is a lot more features in Mist and some texts are not translated to English but fixing of those is still much smaller task than get Mist fork working properly.

or some reason I didn't liked it too much but probably will go test it again

EDIT: Now I remember. It is because of features like SSO and offline installation are only available with subscription.

One option however would be to use API from Cloudpods (which seems to be written with Go) and UI component from Mist. Then in theory it should be possible to have very simple tool to manage at least virtual machines in all platforms.

[olljanat]

https://cloudforet.io/ (This seems to be even part of Linux Foundation)

Seems that this does not even try implement things like server provisioning but instead of it is centralized inventory and easy access to existing resources. But it has very detailed RBAC and SSO which seems to be available for everyone without any subscription (even though there is some subscription available).

So combination of Cloudpods for admins and Cloudforet for end users/management probably would be working combination for many.

One option however would be to use API from Cloudpods (which seems to be written with Go) and UI component from Mist. Then in theory it should be possible to have very simple tool to manage at least virtual machines in all platforms.

But perhaps I will investigate this option a bit more some day.

[Ben0p]

At this stage I still intend to work on mist even if no one else wants to contribute. I am struggling to find time is my only problem.

When I was doing research, I told ChatGPT all my requirements, intentions and what I was trying to achieve. Using the deep research feature it essentially came up with mist. Not saying ChatGPT is entirely accurate and there aren't other solutions out there.

Originally I was going with Apache Cloudstack, but that is more for private cloud infrastructure management, not so much for managing mutli-cloud vendors.

My end game is having a platform to deploy apps and VMs from templates, similar to https://elest.io/, with the option for deploying to cloud or self-hosted infrastructure.

Anyway, it is the start of winter tomorrow in the southern hemisphere, so I should get more time on those rainy days in the coming months to get stuck in to it, hopefully.

[olljanat]

At this stage I still intend to work on mist even if no one else wants to contribute.

Ok, I sent you request to become member with owner rights to mistcommunity org so you are able to continue work in here. Feel free to add more repos, unarchive those, etc what ever is needed.

Biggest challenge in my opinion is that apache/libcloud isn't at least anymore very active project. Example my simple PR apache/libcloud#2073 have been there now two weeks without comments.

In other side version libcloud which Mist-CE is using is very heavily customized and there is quite a lot of Cloud specific expections like this in code:
https://github.com/mistio/mist.api/blob/dcf483f47e7433d7f8bb44eb784a677ce057641f/src/mist/api/clouds/controllers/network/base.py#L447-L452

When I was doing research, I told ChatGPT all my requirements, intentions and what I was trying to achieve. Using the deep research feature it essentially came up with mist. Not saying ChatGPT is entirely accurate and there aren't other solutions out there.

I'm pretty sure that it is quite accurate based on my study too. There is many proprietary multicloud management tools but not many open source.

My end game is having a platform to deploy apps and VMs from templates, similar to https://elest.io/, with the option for deploying to cloud or self-hosted infrastructure.

You might want to check this new API which I build on top of libraries from Cloudpods. Features mentioned in readme works with Aws, Azure and Proxmox already and best part is that very little code is needed for achieve that, thanks to CloudMux library which seems be quite good quality code and which seems to be getting new features like this yunionio/cloudmux#1227 (SUSE Harvester support)

My focus in is VMs only and I want to build API (and later UI) with zero Cloud specific expectations. That why the hardest part actually is figure out that which features actually works in all private and public clouds.

Also I want that it will be possible to use this with delegated rights because in big organizations it is very hard get common tooling agreed. At least Azure part seems to be working fine in that part too, permissions to resource group containing VMs is enough. That why I actually call them for environments, not clouds like Mist does.

Nice part in that approach is that VMs can be deployed without need tell anything about networks, etc.

Anyway, it is the start of winter tomorrow in the southern hemisphere, so I should get more time on those rainy days in the coming months to get stuck in to it, hopefully.

Have fun ;) Here in Europe summer is just starting and I actually will start my vacation in Friday so will try stay offline couple of weeks.
After my vacation will most likely try make cloud-vm-api working with Nutanix and VMware too and then perhaps build either Terraform plugin on top of that API or use Claude to create UI on top of it (assuming of course that I don't get some new more interesting idea to work with...)