Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


A new VPS control panel?
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

A new VPS control panel?

s_ks_k Member
edited April 2013 in General

Hey everyone,

I've been working for a little while on a proof of concept for a new VPS control panel. It's been working quite well, so I'm considering continuing development on it and making it a real product.

Obviously part of my consideration is what the community thinks, so I wanted to ask everyone: what do you think? Do you see a need for a new VPS manager? If so, what do you think it would need to have? What does SolusVM and HyperVM do wrong?

I'm especially interested in hearing the opinions of providers. What frustrations do you have with the software you currently use? What would make you switch? I'd like to know that there's interest on your end before I continue with this.

I've taken some inspiration from joepie91's CVM, but really my program is completely different. It's written in Python, and uses libvirt, which allows it to support many platforms (right now, I'm targeting OpenVZ, Xen, and KVM).

If there's interest, I can share some more technical details and thoughts. I'd love to talk, so feel free to send me a message. If any providers are interested in speaking, let me know. Early adopters would have the potential to be a big influence in its direction.

  • s_k

Comments

  • No

  • Well, show us some screenshots, examples and things first?

  • MaouniqueMaounique Host Rep, Veteran
    edited April 2013

    Hello !

    Features needed:
    Correct pooling and managing of IPv6, similar to IPv4. Assign by various criteria, such as based on IPv4, node, even customer.
    Internal IPv4 managing (RFC 1918)
    Automation for at least WHMCS.
    Consoles for all the virtualizations.
    Customer self-service (tun/tap, fuse, hostname, password, reinstall templates, repartitioning at least for ext3 gpt mbr with data preservation-shrinkin-enlarging for KVM/Xen)
    Veth for OVZ support.
    Separate partition support for xen/kvm, i.e. swap partition separately.
    Customer managed install media (upload of own iso/tar of ovz/xen template.
    I can think of more if needed, but fix these for now :D

  • /etc/vz/dists could use syncronization.

  • jarjar Patron Provider, Top Host, Veteran
    edited April 2013

    @s_k said: What does SolusVM and HyperVM do wrong?

    If I'm correct, you still cannot input difficult passwords in SolusVM. Now, I know that the best security measure is not to change root in SolusVM, but changing it in the panel with a difficult password is still a greater security measure than changing it in the panel with a simple password. If they were holding back for that reason, they'd remove the option. The fact that they care a total of zero about this but have time to play with bootstrap, there's your answer. They do good work, don't get me wrong. They helped us out a lot when we were using them, credit where it's due. Doing a lot of things great doesn't excuse dropping the ball. I'm sure they'll read this and agree with me. There's no bad feelings here, I stopped using them for reasons completely unrelated to quality of their work.

    HyperVM is dead. Only BlueVM can make that beast work worth using these days. It's too bulky for most people to be interested in attempting to update it.

    I want a control panel that works, supports client/slave relationship, is simplistic in nature, isn't ioncube encoded, and...that's about it. I can do the rest :P

    Putting weight (too bad it's not much weight) behind CVM, but always interested in seeing this particular market expand. Given how theoretically simple it should be for even an intermediate programmer (I know @joepie91 is a great coder, and he's going for above simply adequate), I'm surprised there aren't more around. I know @curtisg has done some work but I've never seen the code.

  • joepie91joepie91 Member, Patron Provider

    @s_k said: and uses libvirt,

    Take into consideration that libvirt support for OpenVZ is abysmal. It can fail at any point for no apparent reason.

  • MaouniqueMaounique Host Rep, Veteran
    edited April 2013

    @joepie91 said: It can fail at any point for no apparent reason.

    How I love those fails... In any app, exit 0, but didnt do anything !
    Or, there has been an error, giving up :) Nothing in logs. Good luck :P

  • twaintwain Member

    @jarland - what do you use now?

  • jarjar Patron Provider, Top Host, Veteran

    @twain said: what do you use now?

    VirtPanel. They've been extremely open to development suggestions, and implemented almost everything we've asked for. They won't be our final panel, but I hope that we'll have done them good for the future by working with them to get them where they need to be to get business on this side of the market.

  • marcmmarcm Member

    @jarland - What are you guys using as your VPS control panel these days?

  • jarjar Patron Provider, Top Host, Veteran

    @marcm said: What are you guys using as your VPS control panel these days?

    VirtPanel. I wanted to make it clear that I don't need easy mode, and give clients a different experience. But...ready to settle into something I can grow with.

  • krokro Member

    Use native tool stacks, death to libvirt :)

  • @jarland said: I know @curtisg has done some work but I've never seen the code.

    I've got old archives of my VPS control panel, billing panels, nginx control panels, etc. on my NAS if you ever want to see them.

  • s_ks_k Member

    Was it something I said?

    @concerto49 said: Well, show us some screenshots, examples and things first?

    I've been working on the backend features, the design is currently nonexistent. In the next few weeks I'll work on getting a demo on Youtube.

    @joepie91 said: Take into consideration that libvirt support for OpenVZ is abysmal. It can fail at any point for no apparent reason.
    @kro said: Use native tool stacks, death to libvirt :)

    I've mostly had success with libvirt thus far. Obviously it doesn't support everything, and there are some things that can be done better or faster with tools like vzctl. In these cases, I use vzctl etc. libvirt is the fallback, and if possible, I use the platform's tools.

    @jarland said: Putting weight (too bad it's not much weight) behind CVM, but always interested in seeing this particular market expand.

    Well, I'll see what I can do about that. :)

    @jarland said: Given how theoretically simple it should be for even an intermediate programmer (I know @joepie91 is a great coder, and he's going for above simply adequate), I'm surprised there aren't more around.

    What I've found is that most of the work comes from supporting all the different platforms (i.e. I can't just write a site that wraps vzctl if I want to support KVM and Xen, etc.) I can't even wrap the "logic" in vzctl because there are lots of things OpenVZ does that KVM doesn't. Another challenge has been getting the networking right. e.g. I spent a fair amount of time on static DHCP for guests.

    @Maounique said: Correct pooling and managing of IPv6, similar to IPv4. Assign by various criteria, such as based on IPv4, node, even customer.

    Internal IPv4 managing (RFC 1918)
    Automation for at least WHMCS.
    Consoles for all the virtualizations.
    Customer self-service (tun/tap, fuse, hostname, password, reinstall templates, repartitioning at least for ext3 gpt mbr with data preservation-shrinkin-enlarging for KVM/Xen)
    Veth for OVZ support.
    Separate partition support for xen/kvm, i.e. swap partition separately.
    Customer managed install media (upload of own iso/tar of ovz/xen template.

    • OpenVZ veth support: Already done in my prototype
    • Private addresses (RFC 1918): Could you elaborate? Are you saying you want support for private networks and NAT?
    • WHMCS: Definitely something I'll be supporting if I want any users :)
    • Consoles: Yep, done in the prototype. I'm also considering adding something like noVNC.
    • Separate swap partitions for Xen/KVM: I haven't implemented support for this yet, but will do so if feasible
    • Customer managed install media: I had considered this, but wasn't sure if any providers would want customers uploading massive images/archives. Is this a common requirement? If so, I'll definitely try to get to it.
    • Self service: Hostname, password, reinstalls (templates or mounting ISOs), are all set. Repartitioning/resizing is not something I've implemented yet.
    • TUN/TAP: haven't done anything with this yet, I'll have to look into it

    Thanks for all the comments :)

  • IshaqIshaq Member

    @s_k said: Was it something I said?

    No, curtis leaves unwanted and useless comments.

  • @Ishaq yep that's because opinions are overrated.

  • Maybe this is not the right place to ask... but what happened to CVM?

  • joepie91joepie91 Member, Patron Provider

    @VirtualSRV said: Maybe this is not the right place to ask... but what happened to CVM?

    I'm still working on it... due to unforeseen circumstances everything has been delayed quite a bit. That should be cleared up in a few days, as I found a place to live :)

  • ryanarpryanarp Member, Patron Provider

    @joepie91 awesome that you found a place to live. Looking forward to seeing your panel in action.

  • Instead of a bunch of people (e.g. @BlueVM, @joepie91, @s_k) working on their own individual panels here and there when they have time, how about a pow-wow where you all can collaborate and mesh the best ideas of each and converge into a single panel? It'd move way faster and be better than any of the individual panels.

  • joepie91joepie91 Member, Patron Provider

    @amhoab said: Instead of a bunch of people (e.g. @BlueVM, @joepie91, @s_k) working on their own individual panels here and there when they have time, how about a pow-wow where you all can collaborate and mesh the best ideas of each and converge into a single panel? It'd move way faster and be better than any of the individual panels.

    While that may sound like a good idea in theory, it's very unlikely to work in practice, because:

    1. Cost: some will want to write a panel and release it for free, others will want to charge people for it. And even if a few people all want to charge for it, they probably have different ideas on what kind of model to use.
    2. License: some will want to write open-source code, others want to keep it proprietary.
    3. Architecture: some will want to use SSH for node communication, others will want to use HTTPS, and so on. There are a lot of differences in software architecture that are not going to be compatible.
    4. Security: some will not really care about security, others will want to make sure they've got everything covered.
    5. ...

    Altogether, it's very unlikely if not impossible that any two projects will be on the same 'wavelength' for each of those things. The problem is that "best ideas" can be interpreted subjectively.

  • For the record, joepie91 and BlueVM are working on the same panel. It's joepie91's project we just keep the boat floating.

Sign In or Register to comment.