Top.Mail.Ru
? ?

Linux.com article on Virtualization

Michael Stutz wrote an article for Linux.com entitled Virtualization begins to materialize in the Linux kernel. In the article he addresses, "Two types of virtualization"... paravirtualization and containers. Our friendly OpenVZ developer Kir Kolyshkin is quoted quite a bit. Obligatory quote:
Biederman says that the first thing everybody had to do was learn how to work together. What helped, he says, was working on first implementing "generally non-controversial" features. "Plus, this cemented the incremental approach," he adds, noting that both Linux-VServer and OpenVZ have already announced that they're using these new features in their software.
The article goes on to discuss IPC virtualization, PID virtualization, and UTS virtualization. We need more articles like this one.

I myself have written an introductory article about OpenVZ for SearchServerVirtualization.com:
Intro to OpenVZ: Part I
Intro to OpenVZ: Part II

More articles to come.

Why Virtuozzo is good for OpenVZ

I'm not a big fan of proprietary software, and this is an official blog of OpenVZ, not Virtuozzo — so I haven't had any plans to tell you anything about the latter. But there is one subject I really want to share with you: why Virtuozzo is good for OpenVZ.

To begin with, Virtuozzo is a commercial product based on OpenVZ software. It adds some more value: features, GUIs, tools (and there is also Virtuozzo for Windblows). The first version of Virtuozzo was released about 6 years ago, so it is not something new. Virtuozzo costs money, and is used by big corporations for mission-critical applications. Virtuozzo customers expect it to be very stable, fast, bug-free, well documented, etc. And to sell the product successfully, those expectations must be met.

But how can all this be achieved? A lot of approaches are used: from having a top-notch development team to dedicating lots of resources to testing the software. Indeed, the Virtuozzo QA team is quite large (as well as their server farm), and there are quite a few talented people (and high-end servers) working there. Those people are doing all the bad things with the kernel, trying hard to bring it down to its knees. The majority of the tests are automated, and more and more cases are added to the extensive test suite. The general idea is to find each and every bug — and fix it before the software ships to the customer.

So, good for them, but how does that affect OpenVZ? Quite simple: OpenVZ is the core of Virtuozzo, and thus it is tested and exercised in the same manner. Not too many open source projects can boast of having such great testing and as a result such high quality.

I have to admit it: this is exactly where proprietary software helps open source software. Without Virtuozzo, OpenVZ quality would definitely be lower. A dedicated quality assurance team is always needed because it's just not appealing for developers to find a new bugs in their code — they'd rather write some more.

PS Certainly the scope of "why Virtuozzo is good for OpenVZ" is much broader than just quality: for example, OpenVZ is also feature-rich because of the same reasons.

LinuxWorld Cologne, Germany

We have just returned back from Germany, where I was presenting OpenVZ at Linux World Cologne. For the first time, we had a separate dedicated booth, and we really enjoyed that.

For those three days, we have met hundreds of people, distributed about 100 of OpenVZ DVDs and booklets, and said about million of words telling and showing people what is OpenVZ, how it can be used, what are pros and cons and so on. Surprisingly, a lot of people found it's just what they need.

We also met a number of existing OpenVZ users — for example, one was using it in a small (25 permanent employees + 25 interns) commercial company, where they need a Samba server, a Collax server, and development server, all of three for some reason requiring three different Linux distributions to run on. So, instead of having three servers, they used OpenVZ to create three VEs. It is a production environment working flawlessly for a few months.

I have also met a student who did his thesis on comparing Xen and OpenVZ performance — the thesis is in German but we might translate a part of it later. If you guessed that OpenVZ outperforms Xen than you are damn right.

Finally, in a true spirit of open source, we helped a Debian OpenVZ user to fix a minor misconfiguration and he was finally able to run OpenVZ on his notebook. He is using OpenVZ as a replacement for UML.

OpenVZ on PPC and SPARC

As you might already know, we have recently ported OpenVZ to PPC64 platform, and that was pretty easy.

Here is the news: we are now porting to SPARC, thanks to Jonathan Kinney, a data systems specialist from Washington, USA. It all started with his question which he posted to forum back in May 2006: “if a Sun Fire T2000 could run OpenVZ?”. The answer at that time was like “no, but if we could have access to that hardware, it is quite possible”.

A few months later, in October, Jonathan returned again to that forum thread and said that he has the hardware and is willing to assist in porting OpenVZ to it. Kirill Korotaev picked it up from there, and just a few minutes ago I heard from him that he now has the kernel, which boots and works. Apparently, the size of patch is just about 1500 lines.

What amazes me in this story is the true cooperation between OpenVZ users and developers for the benefit of the worldwide community.

Tags:

On Monday, 23rd Oct the Systems trade fair in Munich opens its gates. It will last till Friday, 27th. On Wednesday, 25th Oct, there will be a talk about OpenVZ and clustering.

The talk is part of the Open Source Day for SMEs day of the Prospects for Open Source conference held at the Systems.

My talk (in German) is titled "Mit Clustertechnik zu mehr Verfuegbarkeit: Ueberraschend kostenguenstig umgesetzt mit Open Source" and starts at 3pm. It describes how to build an high availability cluster with OpenVZ and refers to the HA cluster with DRBD and Heartbeat HOWTO published at OpenVZ Wiki.

To attend the talks on this day there is a fee of 120 Euros, which includes a ticket for the Systems itself, lunch and coffee breaks. For more details, see www.linux-magazin.de/perspektiveopensource.
Good news for all of us on the virtualization front!

The latest prepatch for the stable Linux kernel tree, 2.6.19-rc1, now includes some pieces of OS-level virtualization from OpenVZ, IBM, and Eric Biederman. Those patches have been sitting in -mm (Andrew Morton’s) tree for some time already, and now, during the “2.6.19 merge window,” Andrew has submitted them to Linus Torvalds. So it’s now a part of “vanilla” Linux, and will be finally released as a part of the 2.6.19 kernel when it is released.

So, what exactly went into the Linux kernel? Essentially, three sets of patches that implementing three features needed for any OS-level virtualization solution. Those are IPC and utsname virtualization, and preparations for pid namespaces - click for detailsCollapse )

I am really happy it is a community work and a community process (like I said before). We see different parties bringing in code and expertize, reviewing each other's code, making suggestions, exchanging ideas and improving things — to everybody's benefit!

These are just the first steps. Much more is needed to have full OS-level virtualization in the mainstream Linux kernel. Don’t worry — we are already working on that. A few days ago Kirill sent another iteration (v5) of beancounter patchset for further review and possible inclusion. Beancounters can be used to implement per-VE limits and guarantees for certain resources such as memory.

Red Hat thanks OpenVZ kernel hacker

A bit of shameless PR for our kernel team leader, Kirill Korotaev:

From RHSA-2006-0617:

Important: kernel security update
[...]
* a flaw in the restore_all code path of the 4/4GB split support of
non-hugemem kernels that allowed a local user to cause a denial of service
(panic) (CVE-2006-2932, Important)
[...]
Red Hat would like to thank Wei Wang of McAfee Avert Labs and Kirill
Korotaev for reporting issues fixed in this erratum.

yum for good

Sometimes I want something, err, more technical to this blog. Today I just can not resist the urge.

By popular demand, we created yum repositories for OpenVZ packages. If you are have yum on your OpenVZ system, now you can install/update OpenVZ very easily. Yum is really great tool; we are already using it for maintaining per-VE packages, now it can also be used to update the OpenVZ installation itself.

First, you need to put openvz.repo file into /etc/yum.repos.d/ directory on your OpenVZ box (if there is no such directory, that means either yum is not installed or too old). Second, if you are using any other kernel than stable, enable the appropriate repository by editing openvz.repo file. Now you are all set.

To update your OpenVZ system, just run yum update. Here is how it looks like on my FC5 desktop:
Read more...Collapse )

Last but not least: yum will use mirrors to download all the required packages.
Stephen Shankland wrote an article for ZDNet AU entitled Linux heavies plan lightweight virtualisation. Obligatory quote:
"It's something that we want to see happen," Red Hat's chief technology officer, Brian Stevens, said in an interview during the LinuxWorld Conference & Expo in San Francisco. Red Hat hasn't decided whether to use OpenVZ or Vserver, he added.

Xen is the priority for RHEL 5, due to arrive at the end of the year, but after that will come containers, Stevens said. "I'm looking at that as a RHEL 6 thing," he said.

Novell, which wants to maintain Suse's reputation as the first place to find advanced new features for Linux, is more eager and is considering adding OpenVZ in Service Pack 1 of SLES 10. "We are still evaluating if this is something we can take into SP1," said Holger Dyroff, vice president of Linux product management.

I'm glad to see Novell and Red Hat finally starting to take process container virtualization seriously. Hey, even the Xen guy is quoted as calling it "fabulous" in the uses where it is appropriate.

I must admit that I was a little suprised to hear the Red Hat guy call it a RHEL 6 thing. I mean, saying that is making no commitment to it at all... since they can't really see past RHEL 5 right now. Perhaps it was just a PR parry to the SUSE guy's comment.

Luckily the OpenVZ folks having it working now... although a few iterations and a bunch more testing is a good idea.

Once the bigger distros have it as part of their stock kernels... I would hope it would lead to the eventual migration into the mainline kernel... although admittedly with a *LOT* of pain and work in the process of providing it in a manner that Linus and company would find acceptable. Putting it in vendor kernels will validate its usefulness and that's what the mainline kernel developers care about, right?

Say, is it "virtualization" or "virtualisation"?

OpenVZ contributions to the kernel

I just found out an interesting fact: from the 20 patches that are supposed to be included into the next stable 2.6.17.y (where y=10) kernel, 6 5 were done by OpenVZ developers.

What it could mean, besides the fact that OpenVZ team is a valueable contributor to the mainstream kernel? It also means we do care much for stability and security of OpenVZ, we do a lot of testing and QA, which is good for OpenVZ kernel, but for the mainstream kernel as well.

In a broader sense, this is a nice example of how collaborative open source development works. A nice example of “everybody wins” strategy. Indeed, in open source everybody wins.

Links to individual patches submitted by OpenVZ peopleCollapse )

Update: our kernel team found a bug in this blog post. Looks like the bug belongs to the infamous "off-by-one" category. :) There are actually 5 patches from OpenVZ developers, not 6 — it's just Greg sent one patch twice and thus I counted it twice. Fixed.

Latest Month

July 2016
S M T W T F S
     12
3456789
10111213141516
17181920212223
24252627282930
31      

Syndicate

RSS Atom

Comments

Powered by LiveJournal.com
Designed by Tiffany Chow