Scott Dowdle's blog
I've been keeping up with the virtualization related developments in the upcoming Fedora 15... but even if I weren't... Fedora offers a fedora-virt-preview repository that makes it easy to ckeck out the new stuff on Fedora 14.
Adding SPICE support to virt-manager is one of the upcoming features in Fedora 15 and as of 2011-03-28 it appears to be 100% done. I decided to use the fedora-virt-preview repository to check it out on my Fedora 14 workstation.
If you aren't familiar with virt-manager, it is the default GUI-based management application for virtual machines on Fedora and Red Hat Enterprise Linux... as well as a few other distros. virt-manager uses libvirt so it can support a number of virtual machine types but it is primarily used for KVM and Xen. I use it with KVM and KVM is the only virtualization product that offers SPICE support currently.
Then I deleted the existing VNC-based Graphics device and added a SPICE server. There are a number of different SPICE related options... what port to run it on... whether only the local machine can access it or if it is accessible remotely... use a password or not. There is a setting for SSL port but I'm not sure if that is actually operational... as I have not figured out the SSL stuff yet.
In virt-manager's preferences they have added a toggle for the default graphic device, either VNC or SPICE. Of course you can always delete one and add the other if the default isn't what you wanted.
I posted a contributed OpenVZ OS Template today. The contributed OS Template is Scientific Linux 6 32 bit and it was contributed by Vic from powerpbx.org (firstname.lastname@example.org).
I asked him to share information about he created it and this is what he replied back with via email:
I have no plans to create a x86_64 version or provide regular updates to the x86 version at this time. The only reason I created the x86 version is because I needed a RHEL (or clone) v6 template for my own use. It is easy enough to update/modify/copy by someone else now that this version is out there.
I created it using this procedure and rsync from VMWARE to OpenVZ. Then I manually went through all the installed packages and took out as much as I could to get the size down. When in doubt I compared to the installed packages in a CentOS 5 template.
Yum would not remove kernel so I had to do a "rpm -e --nodeps kernel"
In the newly rsync'ed OpenVZ container I create a file called "vz.repo" in /etc/yum.repos.d with the following text:
then "yum install vzdev vzdummy-apache vzdummy-jre-el5 vzdummy-kernel-el5"
Could not get "vzdummy-glibc" to work. It caused the template to not load on reboot. Someone smarter than me will have to figure that one out. Perhaps vzdummy-glibc needs to be updated for RHEL 6.
Additional things I ran into that appear to be RHEL v6 specific are as follows.
You must comment out "console" in /etc/init/rc.conf and /etc/init/rcS.conf
You must also delete or rename tty.conf and start-ttys.conf.
Turnout for the meeting tonight was great. There were two new people, the regulars, and a few lost sheep that returned. Let's see if I can remember them all... without some last names.
Bruce Stucker(sp?) (new)
Rob Potter (semi-regular... almost back to regular)
David Eder (regular)
Gary Bummer (regular)
Matt (has attended a few meetings)
Jordan Schatz (fairly new, presenter)
Scott Dowdle (me)
Did I miss anybody? Darn it... I forgot to bring my digital camera so I didn't get any pictures. Please someone next month reply to my meeting announcement with a "Hey Scott.. don't forget your camera!", ok? :)
Jordan's presentation on NoSQL was very technical and in-depth. He crammed a lot of information into it... lots of concepts... went over a large number of existing NoSQL projects... why they exist... what environments they come from... how they differ... strengths and weaknesses. We learned about shards as well as various replication styles. We learned about Amazon's paper on Dynamo. We learned about the CAP theorem and ACID. I found the concepts to be very helpful. We learned a little bit about ERLANG. He talked about JSON. He talked about the Thrift and Protocol Buffer protocols... and the fact that many of the NoSQL databases speak HTTP and can in some cases eliminate the need for a webserver. There were lots of acronyms but Jordan explained all that we asked about. Jordan said he would share with us all of his presentations materials. I look forward to that and will post them on the website... because I plan on going through them... following many of the links and doing some reading.
Update: Here are Jordan's links and notes from his presentation:
Then Jordan was asked (I believe by Rob) what development tools he uses in his web development career? EMACS is a big part of almost everything Jordan does as it provides a usable interface to virtually everything. His favorite distro is Debian stable. He likes PHP, especially the enhancements they have made in the 5.3.x series. He prefers git for source control. He likes the CakePHP framework. He likes JQuery. He uses VirtualBox when he has to fire up other OSes to verify browser compatibility. His preferred NoSQL database is Riak.
Jordan told us how busy with work he has been lately and how he stayed up until 3AM last night working on a critical project that isn't finished yet... and that he will probably be up until 3AM again tonight. We decided to give him a break and end the meeting relatively early (9:20ish) so he could get back to work... and thanked him. I gave him a copy of the three-part PBS series from 1997 entitled Triumph of the Nerds as a gift for presenting... although I don't know when he is going to have time to watch it.
For next month Jordan has volunteered to do another presentation. The potential topics are:
- An Introduction to LISP programming with Racket
- Doing everything with EMACS
- An Introduction to source code management with GIT
Any of those sound good to me. Does anyone have a preference?
I've been keeping up with Fedora 15 development. I installed a nightly build on my wife's dual-boot computer. I setup a Fedora 15 KVM virtual machine in preparation for my remix compose... which isn't quite there yet.
Anyway, I've noticed a few changes that came with some updates yesterday that I wanted to share:
- Fedora 15 appears to have incorporated all of the upstream GNOME 3 changes. The experience is exactly like that from the GNOME 3 live beta based on OpenSUSE
- They added a way in the GNOME 3 Shell System Settings to switch back to the GNOME 2 style desktop
- The GNOME 2 style desktop has been polished up some
- Fedora has added some additional artwork for non-GNOME desktops
The GNOME 2 style fallback desktop in GNOME 3 isn't exactly like the previous GNOME 2.32 desktop but it is fairly close. There are some elements from GNOME Shell present... such as the window styling and decorations (although you DO get the minimize and maximize buttons back). You can place application buttons on the top panel but none are there by default. There isn't a right-click desktop menu and the System Settings are from GNOME 3. Although the fallback desktop mode is a bit different than the older GNOME 2.32 desktop, the changes they have made should go a long way to make GNOME 2 diehards a little happier.
I've been using the GNOME 3 Shell more and I really like the changes they've made since the Fedora 15 Alpha release. These changes include larger icons in the Applications list and auto-managed virtual desktops. Some people call this a "dumbing down" of the interface but I prefer to call it streamlining. If you have a lot of previous GNOME 2.x experience you might perceive it to be counter-intuitive... but give it a little while. The streamlining really makes the new environment easy to learn and use. It is elegant, and as one blogger put it... "it looks expensive".
The GNOME developers have frozen the development for this development cycle and only bugfixes will be accepted. GNOME 3 Shell is finished and I like it. As a long-time KDE user, I'm not sure I'm ready to switch to GNOME 3 exclusively but really have enjoyed testing it out. I still expect there to be quite a bit of backlash against GNOME 3 Shell when it becomes the default desktop in Fedora 15 and probably also in the next major releases of other GNOME-based distros... but I think a lot of people are going to like it too. When you get a chance, give it a try.
2011 marks the 20th anniversary of the creation of the Linux kernel. But when exactly was it created? One of the first milestones we have was the original comp.os.minix newsgroup posting that Linus wrote announcing his unnamed new OS. Here's a recreation of that posting:
From: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)
Date: 25 Aug 91 20:57:08 GMT
Subject: What would you like to see most in minix?
Hello everybody out there using minix - I'm doing a (free) operating system (just a hobby, won't be big and professional like gnu) for 386(486) AT clones. This has been brewing since april, and is starting to get ready. I'd like any feedback on things people like/dislike in minix, as my OS resembles it somewhat (same physical layout of the file-system (due to practical reasons) among other things).
I've currently ported bash(1.08) and gcc(1.40), and things seem to work. This implies that I'll get something practical within a few months, and I'd like to know what features most people would want. Any suggestions are welcome, but I won't promise I'll implement them :-)
PS. Yes - it's free of any minix code, and it has a multi-threaded fs. It is NOT protable (uses 386 task switching etc), and it probably never will support anything other than AT-harddisks, as that's all I have :-(.
So that post was sent out on August 25 but it is obvious that Linux has been around for a while... since April... but when was it started? I'm sure a better history is documented somewhere. I guess I'll break out one of my copies of Just for Fun and see if I can find it.
I'm not even sure at what point working name was changed from Freax to Linux. Note that in Linus' post, he doesn't give it a name. I welcome comments from others on what they remember! I started using Linux a little late myself... in January of 1995.
Oh... and it sounds like a good reason to have a party... but when and where? Anyone?
I've been making a personal Fedora remix for a while now... since Fedora 10. While that might sound hard, thanks to Fedora's livecd-tools package and their livecd-creator script, it is really quite easy. I even made a screencast about it. I recently started making a remix of Scientific Linux 6.0 and wanted to share.
As you may recall, I prefer Fedora on my personal desktops but on servers I prefer Red Hat Enterprise Linux or a RHEL clone. There are actually a few clones to pick from and I've been using CentOS for a number of years. One thing I like about CentOS is that one of its goals is to stay as true to RHEL as possible by attempting to be 100% binary compatible with it, bugs and all. Unfortunately the CentOS developers have gotten somewhat backlogged with the onslaught of RHEL releases over the last few months (6.0, 5.6, and 4.9) and have taken a lot of criticism for release delays as well as falling behind on security updates in the process.
Trying out Clone #2
CentOS is definitely the most widely used RHEL clone with an estimated 6 million users who are eagerly awaiting the releases of CentOS 6.0 and 5.6. I can't really fault the CentOS developers for the delays because they are a completely volunteer organization and do development in their spare time.
Another popular RHEL clone is Scientific Linux (SL) which is put together by a small number (two or three?) of developers who are paid to work on it by Fermi National Accelerator Laboratory and the European Organization for Nuclear Research (CERN). SL's main goal is to produce an enterprise grade Linux distribution to meet the needs of scientists and people working with scientific data. SL strays a bit from the stock RHEL package set by adding some additional science related software including some changes to a few core packages to accommodate additional filesystems (reiserfs and AFS). SL is also known for its additional "tweak" packages that are designed to easily change some of the application default configurations.
The primary reason I had previously avoided SL was because I really did not want to deal with their changes and additions to RHEL. Now I'm giving it a try. What has changed? SL has a fairly public development process. For example, they came out with several alpha and beta releases of SL 6.0 before releasing the final version on March 3. They have adopted several of the Fedora developer tools and have given many public presentations about their development process.
While reading about SL I discovered that with their 6.0 release they have switched to Fedora's livecd-creator for producing their Live media. They have also released the kickstart files they used to build their live media and have quite a bit of documentation including a Create your own SL6 LiveCD page. While CentOS does offer live media, they don't use livecd-creator... and their live media does NOT offer an install option. I certainly hope that changes for CentOS 6.0.
Another thing I learned was that as a result of feedback from their userbase, the SL developers have decided to drop their "tweak" packages with 6.0... at least initially... although they may offer them as an option later for those that want them.
But wait... another long post from me... this time from an email I wrote today to the Fusion Linux mailing list... regarding how Flash 64-bit was broken in Fedora 14 and the arguments around who should fix it and why:
----- Original Message -----
> I am sorry to disagree, Linus does state that it is an hack, but he
> also suggests that it could/should be used, please see comment
> https://bugzilla.redhat.com/show_bug.cgi?id=638477#c38 :
> " The nicest alternative might be to just install that mymemcpy.so
> into the google chrome directory, and add the LD_PRELOAD to the wrapper shell
> script that google chrome already uses for the xdg binaries and the ffmpeg
> And obviously something similar should work for firefox. I just
> happen to use chrome, so I gave the directions (approximate as they were) for the
> thing I tried.
> No guarantees. It was a really quick hack. "
Oh don't worry about disagreeing with me. That is often good and I don't take it personally. :)
Yeah... but did you read comment from Linus later on?
- - - - -
Linus Torvalds 2011-02-21 18:19:22 EST Comment 199
Don't use my workaround: it was a stupid hack to test the bug, and show that "always copy upwards" works better than the crap that is in glibc now.
A much better workaround is likely to just implement memcpy() as memmove() (you can replace the inline asm by that in my preload example if you want to). Once memcpy() isn't small and trivial any more, that's just the right thing to do.
The fact that the glibc people don't do that, and that this hasn't been elevated despite clearly being a big usability problem (normal users SHOULD NOT HAVE TO google bugzillas and play with LD_PRELOAD to have a working system), is just sad.
Quite frankly, there is no reason for the current memcpy() mess. There is no _technical_ reason for it, and there is certainly no usability reason for it. Why the Fedora people don't just fix it, I don't understand. It's a shame and a
The fact that Adobe does something that isn't technically right is no excuse for having a sub-par crap memcpy() implementation.
And how does one raise the priority for a bug in bugzilla, or get it re-assigned to somebody who cares?
- - - - -
Please note that I disagree with Linus on everything but the first half of the first paragraph. :) So did the glibc developers and the Fedora developers. While users want it to work, it isn't Fedora's job to fix Adobe's broken program. Just because the problem didn't show up until after the glibc change doesn't mean the problem wasn't there. It was just luck that it worked to begin with. The glibc change just happened to expose the problem, not create it. Adobe needs to fix their program. Why can't they? They update Flash all the time so getting an update out to users really isn't a problem. They said they have a fix for the issue but it could be months before it gets deployed? Why?
Linus still ignores the direct evidence that the glibc change wasn't supposed to be faster except on lower end CPUs... and his testing is invalid. He blathers on and on... intimidating others. That is his way. That is actually his sense of humor... and he is obviously right much more often than when he is wrong... but this is one of the few times he is wrong. :)
Working around Adobe's problem can be done... but why should we do it? Oh, so it makes our distro look better... and users are happier. Yeah, but look at the crazy mess of a workaround it is. Is every distro supposed to engineer their own fix? How much work is that by how many people? I realize that many have not and may not run into this issue because they use older versions of glibc... but you get my point.
I think it is better to say... "we are aware of this bug and we are waiting for Adobe to fix it" and put the blame where it needs to go... rather than everyone working around Adobe's problem and then having to undo everything after they fix it.
What will be next? How many other closed source, commercial vendors will need to be accommodated in the future? This would set a very bad precedent... and that's why (in my opinion) Fedora didn't go for it... even with Linus breathing down their necks.
Fedora doesn't even ship with Flash (nor Google Chrome). They ship with alternative players and those are not affected. Lots of programs break when libraries change... and if they are in distro then they get fixed. If they are closed, commercial products... and they are slow to change... that just re-enforces our belief that FOSS is a better development model... because it is.
Ok, Fusion Linux DOES ship with Flash... and maybe you guys want to fix it. I haven't really contributed to Fusion Linux other than typing some emails here and there... so my opinion doesn't really matter. Do what you think is best... but I did want to provide some additional background and clarification.
I've been living with the warbly sound on some Flash videos for some time now... and I guess I've gotten used to it. Like I said previously, it just strengthens my desire to consume and promote the use of more non-flash content... like webm and ogv.
I don't want baby users who are pampered away from issues... I'd prefer to grow a community of contributors who can see problems (rather than having them hidden from them)... who work to solve problems rather than work around them. I guess that's part of the reason I'm a Fedora user. :)
> I am sorry that it did not fix for you, but as you can check from the
> bug report it fixed for many others. IMHO and until there is proper
> fix we should try to provide a positive user experience to as mush
> users as we can.
That report has been around for a long time and there have been a ton of updates since then. I don't know if that has a bearing on it not working for me or not. This fix I tried was the patch not the Linus fix. That fix was too much work for me.
> I don't think that the technical argumentation on who is right or
> wrong about the proper fix has any relevance for the end user, also
> I do not have have the technical expertise to debate with you, Linus
> or the glibc maintainers about the change.
> My suggestion was just about delivering a better experience to the
> users, getting broken sound on some flash contents is bad, if we
> could avoid it it would be great.
See my above comments.
I do appreciate you taking the time voice your opinions... because it shows you care... and I definitely want to encourage that! Please do not take anything I've written as a personal attack. I don't claim to be any more right than you... but it is obvious I disagree. Perhaps you'll be comforted in the knowledge that Linus agrees with you... I know I would be. :)
One other thing Linus was wrong on and that was on moving cgroup scheduling policy into the kernel... rather than keeping it in userspace... like the systemd developer explained was the better way. I don't recall what the final outcome of that was.
TYL, Scott Dowdle
Here's a LONG response I wrote to one of the comments to the previously mentioned blog posting (The GNOME war) that I wanted to share here as well.
Novell has already been sold. There is an investigation holding up the original transfer date but it is very likely to go through. There will be changes in Novell. If the company taking it over didn't think they could turn it around, they would not have bought it. Turning it around might end up being breaking up all of the pieces and selling some of them, keeping some of them, and killing some of them. Who knows what will happen? We'll just have to wait.
I'm not sure why you seem to be so unhappy with Red Hat with your "while Red Hat is imposing their rules" comment. What rules are they imposing? Is Red Hat in decline? I have no idea. I can tell you that their quarterly reports have been quite positive since they went public... which is pretty rare for any technology company much less a Linux company.
I see some people complaining about their stock prices and valuation... but what tech company on the stock market isn't overvalued? To me the stock market is fundamentally broken but that is a completely different topic so I'll leave it at that.
Oh wait, let me bring up one example. Apple. A while back I read some report where a leading PC magazine had one of their guys dig deeply into Microsoft's yearly earnings reports. The gist of the article was that the author believed he had discovered that Microsoft had moved around various things in their financial reporting to hide the fact that they had lost 1% of the desktop marketshare. Ok, let's think about that for a second. Who did they lose it to? Let's just say all of the 1% went to Apple. Last I checked, and I haven't checked in a while, Apple was very high up on the stock market. They are seen as the darlings of the tech industry... making the cool products... having the best usability... advertising on US TV (I don't know about the rest of the world) with a budget of tens of millions of US dollars. Sure they sold a ton of iPhones but the Android army has come into being and has hit them hard. The iPad has done quite well (15 million sold(?) with the iPad 2 coming out today)... and no one else in the market seems to have an inroad to significant marketshare in the "tablet" arena (which I don't even believe is a legitimate genre although Miguel de Icaza definitely disagrees with me buying his third iPad today). How are they doing with the desktop OS marketshare? They are supposedly selling lots of laptops... but in one of their best years in recent history... they have managed to pull 1% of the marketshare away from Microsoft? 1%? That's all? Yet they are a darling of the stock market... and the envy of the GNOME and Canonical developers.
Of course if you listen to others, the desktop is dead and there is no reason to care about it anymore... and FOSS developers should start working on cloud apps before it is too late... and some say it is already too late.
Wow, I'm getting off on some tangents. In any event, you can see that the tech industry is a tangled web of twisty little passages... all different. :) (Who knows where geeky reference comes from?)
Also, I'm not sure where people keep coming up with this figure that Ubuntu has "60% of the Linux desktop marketshare". I have no idea if it does or not... but determining that is near to, if not completely, impossible. For the sake of argument, let's say it's true. Have they been able to turn a profit yet? If not, why not? How much of the marketshare will they have to gain BEFORE they can turn a profit? Stupid question. Having marketshare for something that is free doesn't make you successful. In fact, it can be a dead albatrose hanging around a company's neck. That is the situation Red Hat found itself in before it decided to go the enterprise Linux route.
Some say it is because Mr. Shuttleworth, who you obviously see as another legendary hero like many see Steve Jobs, has made it impossible for Canonical to make a profit because he has tried to focus the company in too many directions... meaning that no particular direction gets enough focus to be successful. Others might say that doing that is like throwing many things at the wall and seeing which ones stick. He keeps throwing things, and so far nothing has really stuck... nothing that will make the company profitable. I'm not saying that Canonical won't figure it out. I certainly hope they do... because the Linux market needs more FINANCIAL successes... not less.
If Red Hat and Novell falter (which I don't think is going to happen)... while it might shift some customers over to other companies... it will make Linux seem like a less stable technology platform to pay for and invest in. Red Hat has shown that it can be done by having positive financial statements quarter after quarter... all while releasing everything they do as FOSS... and you want them to be taken down? You do know how much they contribute to the Linux kernel, gcc, GNOME, x.org, etc... right?
I do know that if Novell fails or Red Hat falters... any people they have to shed will most likely be snapped up by other companies. Linux can survive the loss of Linus and Linux can survive the loss of one or more of its major distros... but we'd like to do more than survive.
If you ignore everything else I write, please realize that there is plenty of room for more than one or two successful Linux companies. We all do better, when we all do better. :)
Just because you see Arch and Ubuntu and not much else used in your neck of the woods doesn't mean that is how it is everywhere else. Red Hat and Novell are doing well in the "enterprise" space and CentOS is doing quite well too. Debian also.
I think your supposition that if someone uses distro X in high school and/or college they will refuse to work for a company that doesn't use distro X... is silly. Or maybe you were saying that the company they work for will be forced (somehow) to switch to distro X because that's what their new employee(s) use. Riiiiiiiiight. That might be true for major OSes... but not from one flavor of Linux to another. Distros are 95% the same software and switching between them is not so difficult.
Thanks for the discussion, Scott Dowdle
I saw a posting on Fedora Planet entitled, The GNOME wars and just had to respond. Since I put some effort into my comment, I decided to post it here as well.
Your statements are a gross oversimplification of the situation... specifically with regards to GNOME 3 / Shell and Ubuntu Unity.
To date Canonical still has not learned how to properly collaborate with all of their upstreams. Some they have, some they haven't. GNOME is one that they haven't. It took Novell and Red Hat a while to get it right with GNOME and they made their share of mistakes along the way... or at least that is my understanding. The main problem is that in its dealings with GNOME, Canonical would provide completely done software/libraries without much prior collaboration with the GNOME developers on why the library was needed, what needed to be in it, and if any other already existing libraries could have accommodated some or all of the functionality. Just like with Linux kernel development, the developers prefer to be in the loop on developments and having some input and feedback rather than getting a big code dump out of nowhere.
Did Canonical read into that... that Red Hat, which does employ some of the top tier GNOME developers, was trying to block their code? Maybe they did... who knows. Was Red Hat actually trying to block their code? From the top (Red Hat management), absolutely not. That doesn't mean that one or more developers didn't turn their nose up at Canonical, which is possible... but I strongly doubt it. GNOME is a mature community with a wide range of participation from many companies (including Red Hat) as well as independent developers... and Red Hat does not control GNOME.
What we have here is Canonical wanting to have more control over the things that they care about (usability)... with the GNOME and Canonical developers having clashing differences in design decisions. That's all. While some may have reasons to play it other ways, that doesn't make it true.
I actually WISH there were a "war" between Red Hat and Canonical because that would be mean that Red Hat cared more about the desktop. Fedora cares about the desktop, but Red Hat, not so much. While Ubuntu Server may be becoming more popular on servers, I don't think it has eaten into Red Hat's business too much. Even if it had, and Red Hat was trying to be at "war" with them, I doubt they'd do it through GNOME. Ubuntu Server doesn't even ship with a desktop environment.
Who will win? No idea. I'm not even sure there has to be a winner. I've tried both GNOME 3 Shell (in Fedora 15 Alpha) and Ubuntu Unity (in Ubuntu 11.04 Alpha 3). GNOME 3 Shell seems much more polished and streamlined to me. I still haven't quite figured out Unity. If Unity matures and is liked by enough people, other distros will probably add it as an option. If GNOME 3 Shell does well, perhaps Canonical will change its mind. In any event I don't think we'll be able to tell much from the initial releases of either one. It will take time and a few release iterations for things and users to settle.
Having both, at least for the short term, will be a good thing as each project will work harder to compete with the other. For the long term, I'm not sure.
As always, I appreciate your postings as they make me think... and quite frequently, respond. :)
TYL, Scott Dowdle
FOSDEM is the Free and Open source Software Developers' European Meeting that takes place every year in Brussels Belgium. This year it was held on Feb. 5-6. They do a good job of recording and releasing the recordings of their presentatiosn all under a Creative Commons Share Alike license.
Since they released them in Xvid format and they are quite huge, I decided to re-encode them in webm format at a lower resolution, bitrate, and framerate to make them much smaller and more appropriate for streaming... and posted them to archive.org. I've embedded one of Michael Meek's presentation on LibreOffice below but you can find all of the videos here:
Liberating Open Office Development by Michael Meeks
If you can't see the video in your browser, it probably doesn't support webm yet so as an alternative, download the desired video(s) and play it with your preferred media player. VLC is recommended.