Subject: Re: [vserver] Roadmap and Future ...
From: Herbert Poetzl <herbert@13thfloor.at>
Date: Thu, 26 Feb 2009 19:58:20 +0100

On Thu, Feb 26, 2009 at 05:57:54PM +0000, Ed W wrote:

> Herbert Poetzl wrote:
> > One proposal I have (which was actually tested over
> > the years, and seems to work - at least for me) would
> > be that Companies using Linux-VServer get some kind
> > of 'Service Contract' with a Developer (which is a
> > good thing anyway), ranging between a single hour 
> > per month up to a day per week, for, let's say 100 
> > EUR per hour, which can be consumed or not, where in
> > the latter case, this time will be spent on general
> > Linux-VServer development and testing. (if you like
> > that idea, please contact me and I'm certain we can
> > arrange something)

> I'm a (very small) business using vservers and I am extremely 
> keen to see this project survive because I really don't want to have
> to switch to a different project since I do rather like vserver!! I
> think there are several other businesses using vserver here also who
> have raised their heads in the past (obviously many use the project,
> but don't read the mailing list!)

> I will happily buy a Service Contract.  However, I think you need 
> to do a little more work as to how this is all going to work. 

sure, that was just a rough idea where I know that
it works, and it always is a win/win ...

> In the past I have tried to get a few other people to bulk buy work 
> and the issue I think is that businesses (and individuals) really 
> have to define exactly what they are getting when they spend some 
> money (at least anything over pocket money prices). 

exactly and this complicated the whole efford to 
such degree that I would need a department to
define, organize and finalize such effords

> Also in general it's also a better plan to "pick pockets" for 
> pennies and pickup lots of pennies than it is to look for a few 
> large sponsors. 

I'm not really 'looking' for the large sponsors, on
the contrary, this approach will allow every company
who likes to use Linux-VServer to contribute back
in a quite simple way

> Which is to say that perhaps IBM will buy you tomorrow, but until 
> then my personal pain threshold might be very much somewhere 
> between 100 - 1000EU/year, so you really need more than just *my* 
> money to make this work. 

that's perfectly fine and the benefit here is in
the constant/periodical flow of funding at a level
which doesn't hurt ...

> Please think about what you can "sell" me that I can put down on
> my books and capitalise in some way as "goodwill" or "research" or
> whatever. 

well, a service contract for, let's say 50 EUR per
month would fit your case perfectly fine, and as
Linux-VServer user, the books are fine too, if you
'buy' the 'guarantee' that if something goes wrong,
one of the specialists will be available to look
into it

> Also think about how you can interest the 100x as many
> smaller users to contribute as well as the mega businesses?

scales down and up, no problem there, but I do not
really want to 'sell' anything, I want to get folks
to contribute back on a voluntary but still 'business
friendly' basis

> Also I think this is a good time for vserver to come out and 
> start to look like a really professional project. 

well, given that I can spend some more time on
Linux-VServer in general, that probably can be
arranged, although I would like to see folks which
are doing that for a living to join in instead of
the kernel developer(s) designing web pages :)

> That's not a critique of anything specific, just a call for:
> * Can someone on the list help theme the website and make 
>   it look even more spiffy? 
>   (Click into virtualbox/vmware/xen, etc website and they
>    look pretty professional. Vserver has great content, 
>    but could benefit from a slick front page to sell "the brand")

as long as that doesn't impair functionality or
complicates maintainance, this would be more than
welcome.

> * Some additional high level "What is this project" would 
>   be desirable. 
>   Remember this is for the PHB types, not for "me"

I was under the impression that nowadays wikipedia
is the source for such questions, but yeah, a page
dedicated to that won't hurt I guess

> * Mythtv - http://www.mythtv.org/ - seems to be an example of 
>   a fairly big project which has managed to gather some 
>   community help to get that "corporate look" front website 
>   and still have a decent documentation wiki behind the scenes

> * Documentation path for the newer users to get started.  
>   Which is to say there are some REALLY great docs on the wiki, 
>   but they aren't always joined together in a nice path to 
>   lead the user down the getting started  process.  
>   Again, this hopefully is a volunteer project?

of course, volunteers to clean up the wiki(s) are
always welcome

> If I were "sponsoring" something then I would personally 
> like to see more of the following:

> - List of sub projects on the go and their progress. 
>   (Some ability to vote on their priority would be nice also)

if somebody sets something up to do that, I'll
gladly fill it with information

> - Bug tracker (however, this can be a lot of work to manage,
>   so need to  give some thought on efficient process)

we had a bugzilla for some time, wasn't used and
the maintainance overhead was quite hight, until we
get a huge number of bugs in the near future, I'd
prefer to stick to the bug reporting via ML (there
are templates for that) or via IRC in person

> - Git (or similar distributed VCS repo) which is the primary 
>   source for latest patches (I must not have given this enough 
>   thought because I still can't see why a DVCS of some sort 
>   wouldn't help you do the magic that you do?!)

it's actually quite simple, I have something like
git (just not git itself) which works perfectly fine
for me (because I have been using it for at least
7 years now) and changing to a new repository, like
git/svn/hg only causes lots of work and introduces
subtle errors

I always hear how much Linux-VServer would benefit
from a public git/svn/hg repository, but the fact
that we have such repositories and nobody wants to
actually maintain them is a good indication that
something is wrong with that thought

> - Dedication to pushing parts of vserver "upstream" and 
>   transparency on how that's working out(!)  

the problem is, mainline is not interested ...
we did a major code cleanup, which made some code
parts less readable for the developers, to completely
comply with mainline coding style about two years
ago, but mainline keeps reimplementing Linux-VServer
code (which seems to be the only way to get parts of
it mainlined atm)

>   It might not be easy or even possible, but I think it really 
>   needs to be a goal nonetheless?

personally I don't think so, most importantly as 
merging stuff upstream is a very time consuming and
usually frustrating work with no immediate benefit 
for the user or developer

> - Defined prices for "premium support". 
>   Right now we get such excellent support on the mailing 
>   list that it's hard to imagine what you could do here..?

thanks for the compliment!

> I'm not too sure what I would actually want to "buy" 
> (as opposed to "sponsor") right now.  

> I don't have a regular need for "support", other
> than discovering one off issues that I need help with 
> (likely only at the point of purchase of a new machine
>  / new setup, etc)?

look at it more like an agreement to continuously
contribute (kind of sponsor) to development by
buying time, which can be used to implement new
features (if not actually consumed as support time)

> > I think what we need most in the near future is a
> > new, well tested stable release, which can be adopted
> > by distributions and which will be maintained for
> > a reasonable amount of time, and that's actually what
> > I'm working on whenever time permits.

> Tick - works for me.  Also the more "professional" the whole 
> project looks the more likely distros are to use it. 
> Equally the more distros offer it the more likely it is to snowball
> and increase in popularity, etc

> > New features will mainly include adaptations of 'old'
> > features, lost and forgotten over time, and improve-
> > ments on the mainline integration (like cgroups and
> > such), as well as some virtualizations to simplify
> > handling and make it more guest friendly.

> Can I recommend a "baseline" set of features which are well 
> documented in an objective kind of way.

> So for example if we were to boldly assume one service per 
> vserver then we could have a documented setup per service.  

that is only one of the many ways to use the
Technology, but yes, examples and documentation
in this regard are welcome

> Given that the creation of vservers is largely scriptable, 
> this could also lead to some sort of automated testing for 
> for some of the more problematic features?

yes, we wanted to create a test suite for
automated testing (beyond testme.sh/testfs.sh)
but lack of dedicated time didn't permit it (yet)

> I'm not critiquing the docs here. 
> The point is more that as a technical user I can read the 
> "flower page", figure out what I need to get something done 
> and apply it.
> However, less technical people just want to know, 
> "set the IP here, set the netmask here, start it up by running XYZ"

> > One upcoming mainline feature which I will try to
> > integrate as soon as it is available is the check-
> > pointing and resume functionality.

> Oh, cool!
> 
> On the other hand, can you describe what this does that 
> LVM snapshots don't do? 

save processes (and in the future entire guests)
to disk (while they are running) and resume that
state at any time in the future, on a reasonably
similar hardware ...

> > I also plan to do a web frontend to control basic
> > functionality and a minimalistic host distribution
> > to run Linux-VServer Guests on, but I have no idea 
> > when I will find time for that.

> Sounds like a good project to give some space to on the 
> website and let people contribute?  
> I recently found Balsamiq Mockups - http://www.balsamiq.com/ 
> and quite like it for quick sketches of interfaces.

definitely, but for now it is just an idea ...

> Please definitely consider shoving it up on github or 
> similar though.  I find it *sooo* helpful to be able to 
> fork these web projects and apply my own customisations 
> whilst still being able to track mainline

no problem, util-vserver is maintained in a
repository too, so will any non-kernel project

> > Finally, as usual, input here is welcome, so if you
> > have an isolation feature you want to see really soon,

> For the simple level I use it for, vserver is just pefect 
> already! 

> Only get something out which is "stable" in the sense 
> that you call it that!

well, usually Linux-VServer development or even
experimental releases are as least as stable as
the stable mainline kernel releases, but yeah,
we definitely want to get a feature complete and
well tested 'stable' release out

best,
Herbert

> Thanks for everything vserver!!
> 
> Ed W