Status of OSv

[Waldek Kozaczuk]

Not sure if this group is the right one for these questions but OSv Users group does not seem very active.

1. Would you consider OSv to be in alpha, beta or production ready state of development? 
2. Is it ready for production use? Does anybody use it in production for example on AWS? More specifically I would be interested in running JVM-based and Node.js apps.
3. Has anybody tried to build OSv with Open JDK compact profiles (http://openjdk.java.net/jeps/161)? My motivation is to lower the footprint of OSv image even more. And it would be even more appealing once Java 9 is release with project Jigsaw which will allow custom profiles as I understand.
4. I have noticed that the development pace has waned down since October 2015 based on this github graph https://github.com/cloudius-systems/osv/graphs/contributors. Does it indicate the project has matured since then or does it mean that it is being slowly abandoned? What is the future of this project in context of the MIKELANGELO initiative?

Thanks for your answers in advance,

Waldek

PS. I am very excited about OSv and unikernels in general so I am very interested in it future.

  

[Nadav Har'El]

On Fri, Apr 22, 2016 at 5:50 PM, Waldek Kozaczuk <jwkoz...@gmail.com> wrote:

Not sure if this group is the right one for these questions but OSv Users group does not seem very active.

The "osv-user" mailing list was never actually used (one person wrote to it once, almost two years ago), and all the discussion is happening on "osv-dev", so this type of questions is definitely ok here.

I removed "osv-user" from the google-groups directory, as it is indeed of no use.

1. Would you consider OSv to be in alpha, beta or production ready state of development? 

OSv is considered Beta, but there are users who use it on production.

1. Is it ready for production use? Does anybody use it in production for example on AWS? More specifically I would be interested in running JVM-based and Node.js apps.

Since OSv is open source, we do not really know who is using it and for what. I know that in the past people have been using OSv for all the things you mentioned (AWS, JVM and node.js) but I don't know how much "production" use it actually got. I hope that some other members of this list can share their experience with production use of OSv.

 

1. Has anybody tried to build OSv with Open JDK compact profiles (http://openjdk.java.net/jeps/161)? My motivation is to lower the footprint of OSv image even more. And it would be even more appealing once Java 9 is release with project Jigsaw which will allow custom profiles as I understand.

I am not aware that anyone tried, but I don't see any real reason why this should be particularly difficult - OSv already runs (barring bugs) the entire JVM, so I don't see why it should have any problem running a subset of it.
 

1. I have noticed that the development pace has waned down since October 2015 based on this github graph https://github.com/cloudius-systems/osv/graphs/contributors. Does it indicate the project has matured since then or does it mean that it is being slowly abandoned?

The company which started OSv, Cloudius Systems (now called ScyllaDB) is a small startup company, which changed its focus a year ago from OSv to Seastar (http://www.seastar-project.org/) and ScyllaDB (http://www.scylladb.com/). It doesn't mean that we completely abandoned OSv, but we definitely slowed down our development effort on it. At some point we had more than 5 developers working full time on OSv, but today this is no longer the case. We are still doing some work (bug fixes, adding new features such as NFS, etc.) but not as much as we did in the past.

Additionally, OSv is indeed more mature than it used to be. It still has bugs and a lot of room for improvement (unfortunately), but it can run a lot more Linux software than it could in its early days.

 

1. What is the future of this project in context of the MIKELANGELO initiative?

ScyllaDB is indeed a partner of the MIKELANGELO research project funded by the European Union (see https://www.mikelangelo-project.eu/). Among our (ScyllaDB's) other contributions to this project, the project uses OSv as the guest operating system so we have a commitment to ensure that OSv continues to live on.

Nadav.

[kucera...@gmail.com]

sucks that the most usable unikernel implementation (especially with proxmox for example) was abandoned. all the github apps for osv and capstan etc.

I guess it's a containers-only world then...

proxmox could've managed osv deployments just fine.

[Dor Laor]

It's not abandoned, Mikelangelo is active and Waldek himself recently contributed Azure support

and a gui too. That said, containers are eating the world..

--
You received this message because you are subscribed to the Google Groups "OSv Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osv-dev+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[kucera...@gmail.com]

Thank you sir.

"psssst OSv"...
https://news.ycombinator.com/item?id=15008636

On Friday, August 11, 2017 at 12:52:58 PM UTC-4, דור לאור wrote:
> It's not abandoned, Mikelangelo is active and Waldek himself recently contributed Azure support
> and a gui too. That said, containers are eating the world..
>
>
> On Fri, Aug 11, 2017 at 7:48 AM, <kucera...@gmail.com> wrote:
> sucks that the most usable unikernel implementation (especially with proxmox for example) was abandoned.   all the github apps for osv and capstan etc.
>
>
>
> I guess it's a containers-only world then...
>
>
>
> proxmox could've managed osv deployments just fine.
>
>
>
>
>
> --
>
> You received this message because you are subscribed to the Google Groups "OSv Development" group.
>

> To unsubscribe from this group and stop receiving emails from it, send an email to osv-dev+u...@googlegroups.com.

[Waldek Kozaczuk]

I see two reasons unikernels have not gained traction:

• big educational gap - very few people are aware of unikernels and even fewer understand what they are possibly due to lack of some kind of "success story"
  
• unikernels are inherently difficult to use/build which I think is greatly improved by new capstan from MikelAngelo (https://github.com/mikelangelo-project/capstan) that allows composing OSv images from packages at file system level
• another project is UniK (https://github.com/cf-unik/unik)

I myself wrote an article on LinkedIn about OSv (and how great it is) and planing to continue writing about it - next blog will be about composing OSv images using new capstan.

Waldek

[Waldek Kozaczuk]

Forgot to add the link to my article - https://www.linkedin.com/pulse/osv-universal-unikernel-cloud-waldek-kozaczuk

[kucera...@gmail.com]

Nice article! Liked it... didn't know the filesystem was ZFS. The point about removing layers of complexity is a good one.

In particular, lately if you look at the disarray in the Docker community/ecosystem, even docker hub is threatened by this churn, the container format is splitting into alternatives, and no simple non-overkill platform has emerged for mere mortals to use, alternatives like OSv will continue to be looked at particularly if they are usable, and manageable with ordinary open source software for ordinary use cases such as proxmox/kvm.