Virtuozzo and OpenVZ kernel differences [message #10033] |
Tue, 06 February 2007 04:57 |
mikov
Messages: 6 Registered: February 2007
|
Junior Member |
|
|
I am sorry if this question has already been asked.
Are there any kernel differences between OpenVZ and Virtuozzo ?
The page http://openvz.org/support says that Virtuozzo supports higher VPS density, which to me implies functionality implemented in the kernel (but I may very well be wrong).
If there are differences, shouldn't the kernel parts of Virtuozzo be released, since they fall under the GPL ?
Again, I am sorry if this has been asked or is something obvious.
|
|
|
|
|
|
Re: Virtuozzo and OpenVZ kernel differences [message #10060 is a reply to message #10056] |
Wed, 07 February 2007 00:38 |
mikov
Messages: 6 Registered: February 2007
|
Junior Member |
|
|
Thank you for your reply. Please, do not take this as trolling or flaming. I am trying to clarify some issues with respect to Virtuozzo and OpenVZ for a client. I have been told that Virtuozzo probably violates the GPL, and from your reply it appears to be true (in case I misunderstood, I apologize).
dim wrote on Tue, 06 February 2007 11:47 | 1) Kernel modules are not necessarily available under GPL, there are a number of binary modules for Linux - see ati drivers, for example. As I know, Linus doesn't prohibit them.
|
To the best of my knowledge this is not true. The consensus is that only kernel modules which have been developed for another OS and thus cannot be said to "derived" from the kernel _may be_ acceptable. Virtuozzo's kernel code however is obviously derived from the kernel.
Quote: | 2) Open-source licensees and GPL in particular require, that source code should be available with binaries. So, if you are Virtuozzo customer, ask your contact person for sources of these modules.
|
I am sorry but this is clearly wrong. The GPL requires that the source be made available to everybody, not only customers.
|
|
|
|
Re: Virtuozzo and OpenVZ kernel differences [message #10069 is a reply to message #10060] |
Wed, 07 February 2007 11:23 |
dim
Messages: 344 Registered: August 2005
|
Senior Member |
|
|
mikov wrote on Tue, 06 February 2007 19:38 |
To the best of my knowledge this is not true. The consensus is that only kernel modules which have been developed for another OS and thus cannot be said to "derived" from the kernel _may be_ acceptable. Virtuozzo's kernel code however is obviously derived from the kernel.
|
You are mistaken. Virtuozzo is available for at least 2 OSs - Linux and Windows, and there were also FreeBSD port. So, mentioned modules are covered by this consensus. But anyway, my point of view is not official SWsoft's point of view, because I'm even don't know, which license these modules are covered by, when provided to customers. I want just say, that it doesn't matter, because they _may be_ covered by any license.
|
|
|
|
|
Re: Virtuozzo and OpenVZ kernel differences [message #10717 is a reply to message #10651] |
Tue, 27 February 2007 03:15 |
mikov
Messages: 6 Registered: February 2007
|
Junior Member |
|
|
If it is only the user mode tools, they can put whatever restrictions on them. Not good for free software, but SWSoft must make money after all.
However I am convinced that all virtualization related kernel extensions are inherently based on the Linux kernel, and so can only be licensed under the GPL. This is what my initial post was about.
By the terms of the GPL, SWSoft MUST provide their customers with the source of the kernel modules at the time of the sale; the customers MUST be free to do anything with the source - any additional restrictions would violate the GPL. I am assuming they are doing this. Alternatively SWSoft MUST make their kernel changes publicly available on their web site - apparently they are not doing that.
If they are not doing that either one of these things, they are probably in a violation of the GPL. However I personally have no great desire to stir this too much, because they are making a great contribution by maintaining a GPL OpenVZ. So they are _the good guys_ even if they are not 100% perfect - nobody is.
|
|
|