SOLUSVM acquired by PLESK

LosPollosHermanos

@Francisco said:

Hxxx said: You mean re-base it? Put some laravel on that shiiii if you are staying with PHP.

Are you going full Angular 4/5/6 ? Vue? React? Or staying with the good ol jQuery?

Laravel is what we'll rebase to I've had to write a lot of things for Codeigniter that doesn't really work nicely just because of how CI works...and then laravel has it built in anyway. Once I'm done my tour in Vegas and head home i'll get block storage fleshed out in the current code base then likely start the rebasing soon after.

As for the front end, probably still JQuery. I'm not quite webscale enough to dabble in Angular.

My main issue is going to be finding a UX designer. If I get a good starting block I can usually build pages from there (like I did on the current stallion) but I'm not creative enough to do it from scratch.

Francisco

I made my own cloud control panel with Laravel + AdminLte for application layer. Works well. I still use Solus on the back end though for all the virtualization stuff.

Francisco

LosPollosHermanos said: You can't go wrong with Laravel. As long as you stick with the core stuff and don't go off on tangents using the latest shiny objects that are all the rage today and then all of a sudden something else comes along that newer and shinier.

I usually try to keep things very simple.

Stallion as a whole is only like 60k lines of code including all the frontend (js, css, html). Moving to laravel will chop that by 20k just because i'm not having to reinvent the wheel.

Francisco

Hxxx

Yeah that's what I like of Laravel. I'm not yet a guru on it but I'm moving that way. Got tired of doing too many steps that are already done in a Framework such as Laravel and their Eloquent.

It has so many things already done. Authentication? php artisan make:auth , Done.

Slim is also Gorgeous for Rest API. For those looking for simplicity.

@Francisco said:

LosPollosHermanos said: You can't go wrong with Laravel. As long as you stick with the core stuff and don't go off on tangents using the latest shiny objects that are all the rage today and then all of a sudden something else comes along that newer and shinier.

I usually try to keep things very simple.

Stallion as a whole is only like 60k lines of code including all the frontend (js, css, html). Moving to laravel will chop that by 20k just because i'm not having to reinvent the wheel.

Francisco

joepie91

@MikePT said:

@Hxxx said:
Solus is basically abandoned. The code is probably so terrible that OnApp didn't really wanted to push forward with it.

This is the time for the panties companies to actually put some effort and do some in-house development. Or just go the ez route and buy some Virtualizor until the history repeat.

I doubt the code is terrible.

I've seen the code. I can assure you that the SolusVM code is some of the worst code I've ever seen. And that says a lot.

One particular example that stayed with me, was the massive abuse of execs all over the place; invoking complex chains of command-line tools (with a bunch of RCE vulnerabilities due to non-existent safety measures) for implementing functionality that could have trivially been - safely! - implemented in PHP directly.

SolusVM is, or at least was, quite literally a graphical shell around a bunch of command-line tools. PHP was just used as glue; much of the business logic wasn't even written in PHP.

@AnthonySmith said:

KuJoe said: Anybody with a background in coding or who's more intelligent than me (not hard) could easily knock out something equal to or better than SolusVM in less time it takes SolusVM to release an update.

yep, I understand your perspective, its just hard for me to accept that if that (quoted) was the case it would not have happened already.

Perhaps I am naive.

It's easy to build a VPS control panel. It's much harder to build a generic panel that can also deal with other people's edge cases. It's even harder to actually get other companies to adopt it, in the current fucked-up climate. The problem isn't the technical aspect, it's the non-technical aspects; the fuzzy human factor.

This is most likely why it was simple for @KuJoe; he only needed it to support his own business and usecases, and of course already convinced himself to use it.

AnthonySmith

yep, well that's apples and oranges, a VPS control panel, probably not so hard, a multi essentially anonymous user front end with API billing panel integrations ip management, 1 click functions for end users, migration elements for admins (I could go on) specifically suited to the hosting industry and then making it possible to import from other panels (the reason no one has dethroned solusvm to date despite the financial incentives) is another case entirley, and not something I believe can be done in 40, 50, 100 or even 400 hours.

Francisco

MikePT said: I doubt the code is terrible.

Just as an FYI, but when I did my audit years ago this was a fun case I ran into.

They have their solusvm.conf file which stores the mysql username/password/database/salt, it's separated by a colon. The smart idea would be to instead just have a $config array and then they can just grab out the keys, but whatever, splitting a string is "fine".

The big problem is they didn't use explode() to do the splitting, they literally broke out to shell with shell_exec and called cat, cut, and awk for each field. This means this bash monster was called 5 times to get 5 different fields.

Instead of just wrapping CURL they have/had this multi thousand line file that "implements" an HTTP request wrapper using fopensocket and parsing.

Francisco

letbox

Solus not that good of coding we even do better with our vPanel control however since OpenVZ now EOl we stopped to developing. We may take over with KvM with our panel but I double will be soon since our coder very busy.

raindog308

Francisco said: My main issue is going to be finding a UX designer. If I get a good starting block I can usually build pages from there (like I did on the current stallion) but I'm not creative enough to do it from scratch.

Stallion looks hella better than Solus.

Francisco

raindog308 said: Stallion looks hella better than Solus.

Current stallion is nice...and also built on bootstrap 2.0.

I need at the very least to bring it up to whatever the latest bootstrap is :P

Francisco

jetchirag

@Francisco CI is pretty easy to pick up for newbies but laravel is indeed a beast. If you are going to be rewriting, why not python/Django?

Francisco

jetchirag said: @Francisco CI is pretty easy to pick up for newbies but laravel is indeed a beast. If you are going to be rewriting, why not python/Django?

Because i'm not a masochist.

My python experience is very limited so i'd rather stick with what I know for core/production stuff :P

Anyway, this is going off topic. Back to PleskVM.

Francisco

KuJoe

@AnthonySmith said:
yep, well that's apples and oranges, a VPS control panel, probably not so hard, a multi essentially anonymous user front end with API billing panel integrations ip management, 1 click functions for end users, migration elements for admins (I could go on) specifically suited to the hosting industry and then making it possible to import from other panels (the reason no one has dethroned solusvm to date despite the financial incentives) is another case entirley, and not something I believe can be done in 40, 50, 100 or even 400 hours.

Most of that can be done pretty easy, the only major issue would be the import script because SolusVM's database is all over the place (for me it was easier to just use their database scheme as the foundation rather than converting it).

This thread makes me want to build a new open source control panel from scratch now, anybody want to build a theme?

AnthonySmith

should I rename this thread, "replacing solusvm is easy - see you in 4 years for the same discussion with no action?" haha. (been guilty of it myself)

MikeA

@AnthonySmith said:
should I rename this thread, "replacing solusvm is easy - see you in 4 years for the same discussion with no action?" haha.

From Plesk's blog post, they seem to want to do good with it, and I would guess this gives Phill more room to do what he wants since he's with the Plesk team now.

Plesk is acquiring all SolusVM assets and will carry on its growth strategy. Thus, elevating it from a single server control panel to a future-proof cloud platform. We’ll provide both cloud service providers and web professionals a single pane of glass application that will become the future of WebOps.

AnthonySmith

MikeA said: From Plesk's blog post, they seem to want to do good with it, and I would guess this gives Phill more room to do what he wants since he's with the Plesk team now.

I think I am probably up there as one of their biggest critics, its no secret, but this time, I really do believe we will see some good stuff come out of solusvm.

Also this time I do not expect the price to stay the same, if things improve I don't mind paying more and I don't expect PLESK to swallow the VAT as OnApp did, it was quite an unusual move to begin with and for the past 4 years.

MikePT

@joepie91 said:

@MikePT said:

@Hxxx said:
Solus is basically abandoned. The code is probably so terrible that OnApp didn't really wanted to push forward with it.

This is the time for the panties companies to actually put some effort and do some in-house development. Or just go the ez route and buy some Virtualizor until the history repeat.

I doubt the code is terrible.

I've seen the code. I can assure you that the SolusVM code is some of the worst code I've ever seen. And that says a lot.

One particular example that stayed with me, was the massive abuse of execs all over the place; invoking complex chains of command-line tools (with a bunch of RCE vulnerabilities due to non-existent safety measures) for implementing functionality that could have trivially been - safely! - implemented in PHP directly.

SolusVM is, or at least was, quite literally a graphical shell around a bunch of command-line tools. PHP was just used as glue; much of the business logic wasn't even written in PHP.

@AnthonySmith said:

KuJoe said: Anybody with a background in coding or who's more intelligent than me (not hard) could easily knock out something equal to or better than SolusVM in less time it takes SolusVM to release an update.

yep, I understand your perspective, its just hard for me to accept that if that (quoted) was the case it would not have happened already.

Perhaps I am naive.

It's easy to build a VPS control panel. It's much harder to build a generic panel that can also deal with other people's edge cases. It's even harder to actually get other companies to adopt it, in the current fucked-up climate. The problem isn't the technical aspect, it's the non-technical aspects; the fuzzy human factor.

This is most likely why it was simple for @KuJoe; he only needed it to support his own business and usecases, and of course already convinced himself to use it.

So SolusVM code sucks. But no one else does it better. Other than Virtualizor that seem to develop a lot of features though its just buggy.

If its so easy then do it. Market needs it. Bring some competition and it will make all similar products to evolve as well.
I trust Phill's code and I like him. Keep in mind that he was the main dev and owner but was not the only dev. Though most stuff was done by him.

Anyway its easy to criticize, but what about doing it better? After all the shitty code got them thousands of monthly licenses. Profit.

imok

Francisco said: My main issue is going to be finding a UX designer

I'm not the best UI/UX designer of the LET world. But I'm sure I can help.

classy

@Francisco said:

Hxxx said: You mean re-base it? Put some laravel on that shiiii if you are staying with PHP.

Are you going full Angular 4/5/6 ? Vue? React? Or staying with the good ol jQuery?

Laravel is what we'll rebase to I've had to write a lot of things for Codeigniter that doesn't really work nicely just because of how CI works...and then laravel has it built in anyway. Once I'm done my tour in Vegas and head home i'll get block storage fleshed out in the current code base then likely start the rebasing soon after.

As for the front end, probably still JQuery. I'm not quite webscale enough to dabble in Angular.

My main issue is going to be finding a UX designer. If I get a good starting block I can usually build pages from there (like I did on the current stallion) but I'm not creative enough to do it from scratch.

Francisco

@KuJoe said:

@AnthonySmith said:
yep, well that's apples and oranges, a VPS control panel, probably not so hard, a multi essentially anonymous user front end with API billing panel integrations ip management, 1 click functions for end users, migration elements for admins (I could go on) specifically suited to the hosting industry and then making it possible to import from other panels (the reason no one has dethroned solusvm to date despite the financial incentives) is another case entirley, and not something I believe can be done in 40, 50, 100 or even 400 hours.

Most of that can be done pretty easy, the only major issue would be the import script because SolusVM's database is all over the place (for me it was easier to just use their database scheme as the foundation rather than converting it).

This thread makes me want to build a new open source control panel from scratch now, anybody want to build a theme?

I’m integrating https://github.com/tabler/tabler into my Laravel based SaaS, maybe something for you guys to use as well?

Francisco

That's really cool!

Thank you!

Francisco

joepie91

@MikePT said:

@joepie91 said:

@MikePT said:

@Hxxx said:
Solus is basically abandoned. The code is probably so terrible that OnApp didn't really wanted to push forward with it.

This is the time for the panties companies to actually put some effort and do some in-house development. Or just go the ez route and buy some Virtualizor until the history repeat.

I doubt the code is terrible.

I've seen the code. I can assure you that the SolusVM code is some of the worst code I've ever seen. And that says a lot.

One particular example that stayed with me, was the massive abuse of execs all over the place; invoking complex chains of command-line tools (with a bunch of RCE vulnerabilities due to non-existent safety measures) for implementing functionality that could have trivially been - safely! - implemented in PHP directly.

SolusVM is, or at least was, quite literally a graphical shell around a bunch of command-line tools. PHP was just used as glue; much of the business logic wasn't even written in PHP.

@AnthonySmith said:

KuJoe said: Anybody with a background in coding or who's more intelligent than me (not hard) could easily knock out something equal to or better than SolusVM in less time it takes SolusVM to release an update.

yep, I understand your perspective, its just hard for me to accept that if that (quoted) was the case it would not have happened already.

Perhaps I am naive.

It's easy to build a VPS control panel. It's much harder to build a generic panel that can also deal with other people's edge cases. It's even harder to actually get other companies to adopt it, in the current fucked-up climate. The problem isn't the technical aspect, it's the non-technical aspects; the fuzzy human factor.

This is most likely why it was simple for @KuJoe; he only needed it to support his own business and usecases, and of course already convinced himself to use it.

So SolusVM code sucks. But no one else does it better. Other than Virtualizor that seem to develop a lot of features though its just buggy.

If its so easy then do it. Market needs it. Bring some competition and it will make all similar products to evolve as well.
I trust Phill's code and I like him. Keep in mind that he was the main dev and owner but was not the only dev. Though most stuff was done by him.

Anyway its easy to criticize, but what about doing it better? After all the shitty code got them thousands of monthly licenses. Profit.

Because there are thousands of markets that are easier to enter than this one. We're talking essentially critical core infrastructure here, where people would often rather run the known quantity no matter how broken it is, than to switch to something unknown.

If my interest were just in building a profitable software business, "building a VPS panel" would not be what I'd go for. I suspect it's the same for most competent developers.

raindog308

jetchirag said: @Francisco CI is pretty easy to pick up for newbies but laravel is indeed a beast.

When I did more php, I loved CI. It gave a lot without going too heavy on the "thou must". I liked it a lot better than CakePHP.

I keep meaning to read a book on Laravel.

jetchirag said: If you are going to be rewriting, why not python/Django?

Probably because @Francisco knows php and not python, I'm guessing.

Fuck you all...now I have an itch to write a panel.

Didn't @joepie91 start one at one point?

jetchirag

@raindog308 said:

jetchirag said: @Francisco CI is pretty easy to pick up for newbies but laravel is indeed a beast.

When I did more php, I loved CI. It gave a lot without going too heavy on the "thou must". I liked it a lot better than CakePHP.

I keep meaning to read a book on Laravel.

jetchirag said: If you are going to be rewriting, why not python/Django?

Probably because @Francisco knows php and not python, I'm guessing.

Fuck you all...now I have an itch to write a panel.

Didn't @joepie91 start one at one point?

I'd be honest that I get that itch every time someone opens a solusvm thread.

I'm currently writing a pull based backup system. I've named it 'pulley.bak' :P

About UI/UX, @Francisco is there any reason for you wanting custom design? There are so many good themes available (both paid and free)?

Francisco

jetchirag said: About UI/UX, @Francisco is there any reason for you wanting custom design? There are so many good themes available (both paid and free)?

So that we don't look like the rest

The one mentioned above is pretty good. I can likely play with the colours a bit and get a current Stallion feel to it.

That'll be later in the year, but for sure bookmarking it.

Francisco

BharatB

It sure is easy to whip up a panel given security stuff is met and as many told exec is used obnoxiously.

Name of the panel is not yet confirmed but it supports only KVM

http://pulsestats.io/let/login.png

http://pulsestats.io/let/admin-dashboard.png

http://pulsestats.io/let/admin-images.png

http://pulsestats.io/let/admin-instances.png

http://pulsestats.io/let/admin-settings-01.png

http://pulsestats.io/let/admin-settings-02.png

http://pulsestats.io/let/admin-settings-03.png

http://pulsestats.io/let/admin-settings-04.png

http://pulsestats.io/let/admin-settings-05.png

http://pulsestats.io/let/admin-tasks.png

http://pulsestats.io/let/admin-task-log.png

http://pulsestats.io/let/login.png

http://pulsestats.io/let/user-create-instance.png

http://pulsestats.io/let/user-instance-logs.png

http://pulsestats.io/let/user-instance-manage.png

http://pulsestats.io/let/user-keys.png

http://pulsestats.io/let/user-recipes.png

http://pulsestats.io/let/user-profile.png

Levi

BharatB said: Name of the panel is not yet confirmed but it supports only KVM

If it's ain't opensource'd -> GTFO.

AnthonySmith

I think you missed the point of a dashboard when 80% of the space is used up with info on 1 server.

BharatB

@AnthonySmith said:
I think you missed the point of a dashboard when 80% of the space is used up with info on 1 server.

You missed the selection list?

jetchirag

@AnthonySmith said:
I think you missed the point of a dashboard when 80% of the space is used up with info on 1 server.

I actually find it pretty useful given that we hypervisor is selectable

BharatB

The dashboard is far from complete though, billing is pending I have the hourly thing going and working on invoices once invoices are done I'll see how to show data on the dashboard.

For the curious minded its VueJS + Laravel REST API for master & Lumen for slave.

jetchirag

Have you temporarily named it Virtkit?