Yesterday I read that Fedora Legacy has become defunct and I'm guessing that means it won't be pushing any updates anymore. As that stands, it probably means I have even more reasons to upgrade to a newer version. With that, I was wondering if anybody has any inside knowledge of when Red Hat might be releasing RHEL 5 (I know this isn't strictly a Fedora question, but I know there are various Fedora users who also work for RH and might know). Once it's released, I'd like to get the equivalent version of CentOS for my server, but if it's going to be a while (say, longer than a month from now), I'll just upgrade my server to Fedora Core 6 during Christmas break.
Any enlightenment would be helpful.
Thanks, Justin W
RHEL5 is supposed to be released next year.
On 12/19/06, Justin W jlist@jdjlab.com wrote:
Yesterday I read that Fedora Legacy has become defunct and I'm guessing that means it won't be pushing any updates anymore. As that stands, it probably means I have even more reasons to upgrade to a newer version. With that, I was wondering if anybody has any inside knowledge of when Red Hat might be releasing RHEL 5 (I know this isn't strictly a Fedora question, but I know there are various Fedora users who also work for RH and might know). Once it's released, I'd like to get the equivalent version of CentOS for my server, but if it's going to be a while (say, longer than a month from now), I'll just upgrade my server to Fedora Core 6 during Christmas break.
Any enlightenment would be helpful.
I think the question I'm more interested in the answer to is; will somebody be picking up Legacy?
They state "The current model for supporting maintenance distributions is being re-examined." which doesn't actually explain what the problem with the old method was.. People, hardware, hosting, donations??
Legacy is quite handy as we know, and it'll be nice to see it up and running again.
On Wed, 2006-12-20 at 09:18 +0900, Shawn wrote:
Once it's released, I'd like to get the equivalent version of CentOS for my server, but if it's going to be a while (say, longer than a month from now),
CentOS wont be ready for at least a month after RHEL5. Has taken 2-3 months in the past.
Shawn
On Wed, Dec 20, 2006 at 10:19:05AM +0900, Naoki wrote:
I think the question I'm more interested in the answer to is; will somebody be picking up Legacy?
No one with sufficient resources has stepped forward.
They state "The current model for supporting maintenance distributions is being re-examined." which doesn't actually explain what the problem with the old method was.. People, hardware, hosting, donations??
People. Specifically, 1-3 basically full-time positions -- so, money too.
On Tue, 2006-12-19 at 20:27 -0500, Matthew Miller wrote:
On Wed, Dec 20, 2006 at 10:19:05AM +0900, Naoki wrote:
I think the question I'm more interested in the answer to is; will somebody be picking up Legacy?
No one with sufficient resources has stepped forward.
They state "The current model for supporting maintenance distributions is being re-examined." which doesn't actually explain what the problem with the old method was.. People, hardware, hosting, donations??
People. Specifically, 1-3 basically full-time positions -- so, money too.
You nailed that one right on the head. They're spending a lot of effort and the cha-ching to bring FC6 to the table, why not use it? So far, it works like a dream. Plus, the kernel and gcc and all the rest of the Linux world has progressed as well, which is all wrapped up into FC6. Might as well bite the bullet and move upwards and onwards.
Jus' my two cents worth. Ric
What kind of storage requirements are we talking about to host the legacy distros?
- Donald Tripp dtripp@hawaii.edu ---------------------------------------------- HPC Systems Administrator High Performance Computing Center University of Hawai'i at Hilo 200 W. Kawili Street Hilo, Hawaii 96720 http://www.hpc.uhh.hawaii.edu
On Dec 19, 2006, at 3:19 PM, Naoki wrote:
I think the question I'm more interested in the answer to is; will somebody be picking up Legacy?
They state "The current model for supporting maintenance distributions is being re-examined." which doesn't actually explain what the problem with the old method was.. People, hardware, hosting, donations??
Legacy is quite handy as we know, and it'll be nice to see it up and running again.
On Wed, 2006-12-20 at 09:18 +0900, Shawn wrote:
Once it's released, I'd like to get the equivalent version of CentOS for my server, but if it's going to be a while (say, longer than a month from now),
CentOS wont be ready for at least a month after RHEL5. Has taken 2-3 months in the past.
Shawn
-- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
On Tue, Dec 19, 2006 at 03:34:54PM -1000, Donald Tripp wrote:
What kind of storage requirements are we talking about to host the legacy distros?
A few hundred gigabytes. That's really not the issue.
Shawn wrote:
Once it's released, I'd like to get the equivalent version of CentOS for my server, but if it's going to be a while (say, longer than a month from now),
CentOS wont be ready for at least a month after RHEL5. Has taken 2-3 months in the past.
Shawn
Thank you. I finally found an article which said RHEL 5 should be released about mid-February, and with a 2 or 3 month window before CentOS is released, it means it could be May before I can get a hold of it. That very nicely coincides with summer break. I think I'm going to take the time to install FC6 to my server, and that will also coincide pretty closely with a theoretical end of support for FC6 (assuming 6 month life spans).
Again, thanks to all for your input. It's exactly what I needed to know.
Justin
On Tue, 2006-12-19 at 21:57 -0600, Justin W wrote:
Shawn wrote:
Once it's released, I'd like to get the equivalent version of CentOS for my server, but if it's going to be a while (say, longer than a month from now),
CentOS wont be ready for at least a month after RHEL5. Has taken 2-3 months in the past.
Shawn
Thank you. I finally found an article which said RHEL 5 should be released about mid-February, and with a 2 or 3 month window before CentOS is released, it means it could be May before I can get a hold of it. That very nicely coincides with summer break. I think I'm going to take the time to install FC6 to my server, and that will also coincide pretty closely with a theoretical end of support for FC6 (assuming 6 month life spans).
You might want to stick with Fedora as the announces also say that Fedora is to become static and meshed with RHEL in some fashion. Ric
On Wed, Dec 20, 2006 at 03:02:10AM -0500, Ric Moore wrote:
You might want to stick with Fedora as the announces also say that Fedora is to become static and meshed with RHEL in some fashion. Ric
It is already meshed with RHEL in a pretty direct fashion. I don't think it's going to become "static", though, although I'm not exactly sure what you mean by that.
On Wed, 2006-12-20 at 07:43 -0500, Matthew Miller wrote:
On Wed, Dec 20, 2006 at 03:02:10AM -0500, Ric Moore wrote:
You might want to stick with Fedora as the announces also say that Fedora is to become static and meshed with RHEL in some fashion. Ric
It is already meshed with RHEL in a pretty direct fashion. I don't think it's going to become "static", though, although I'm not exactly sure what you mean by that.
That was the announce over a month or so ago. FC6 is it. End of the line of the fast paced devel cycle as I recall. There was a thread on the topic.
<quote> "FC6 was the final release of Fedora Core. A much stronger software distribution and community project will take its place. There has never been a better time to get involved in the Fedora Project. Fedora contributors today shape the direction and accelerate the rate of progress of the entire FOSS ecosystem.
There were many other things in the works discussed, but I for now will take a nap. Others will be blogging soon about these topics, and Greg will be updating the Wiki with details and action items." <end-quote>
http://wtogami.livejournal.com/11707.html
There you have it, Ric
Ric Moore <wayward4now <at> gmail.com> writes:
Fedora Project. Fedora contributors today shape the direction and accelerate the rate of progress of the entire FOSS ecosystem.
^^^^^^^^^^
Does that sound like "End of the line of the fast paced devel cycle" to you? What the lines you quoted are actually about is that Core and Extras are merging. But the result will still be released every 6 months as far as I know. According to a draft plan on the Fedora Advisory Board mailing list, Fedora 7 is being scheduled for late April. This is of course subject to change, but it's not going to change to a slow RHEL-style cycle.
Maybe you read what you wanted to hear, maybe you read what you were afraid of, but either way what you read is not what was being told. Sorry.
Kevin Kofler
Kevin Kofler wrote:
Ric Moore <wayward4now <at> gmail.com> writes:
Fedora Project. Fedora contributors today shape the direction and accelerate the rate of progress of the entire FOSS ecosystem.
^^^^^^^^^^
Does that sound like "End of the line of the fast paced devel cycle" to you? What the lines you quoted are actually about is that Core and Extras are merging. But the result will still be released every 6 months as far as I know. According to a draft plan on the Fedora Advisory Board mailing list, Fedora 7 is being scheduled for late April. This is of course subject to change, but it's not going to change to a slow RHEL-style cycle.
Maybe you read what you wanted to hear, maybe you read what you were afraid of, but either way what you read is not what was being told. Sorry.
Kevin Kofler
Development is pretty much stabilized regarding dependencies. It surely does not seem that the pace has slowed down a great deal. As far as things seem, fc7 is in progress and should be as stated above, April or about.
Legacy, I don't know. Jim
On Thu, 2006-12-21 at 01:33 -0500, Ric Moore wrote:
On Wed, 2006-12-20 at 07:43 -0500, Matthew Miller wrote:
On Wed, Dec 20, 2006 at 03:02:10AM -0500, Ric Moore wrote:
You might want to stick with Fedora as the announces also say that Fedora is to become static and meshed with RHEL in some fashion. Ric
It is already meshed with RHEL in a pretty direct fashion. I don't think it's going to become "static", though, although I'm not exactly sure what you mean by that.
That was the announce over a month or so ago. FC6 is it. End of the line of the fast paced devel cycle as I recall. There was a thread on the topic.
<quote> "FC6 was the final release of Fedora Core. A much stronger software distribution and community project will take its place. There has never been a better time to get involved in the Fedora Project. Fedora contributors today shape the direction and accelerate the rate of progress of the entire FOSS ecosystem.
There were many other things in the works discussed, but I for now will take a nap. Others will be blogging soon about these topics, and Greg will be updating the Wiki with details and action items." <end-quote>
http://wtogami.livejournal.com/11707.html
There you have it, Ric
Hi, Rick, I read the blog, and it said that Fedora Core was going to merge with Fedora Extras (which is the way the DVD is packaged now anyway) and become just Fedora. In other words, the apparent plan is to put to gether the OS, and a suite of tools to make the users computer useful. I think this will make Fedora more attractive, especially if they get the stuff all bound together so that all the chasing down of plugins, dependencies and packages can be done prior to the new package being shipped. Also this should permit tighter control over the entire suite of system and tools, enabling a distribution that will be more or less autonomous and correctly configured at the outset. This will displace windows quite easily especially due to the hassels of getting windows, the support costs, and the licensing fees and so forth that cause windows to be one of the most hated OS's ever produced.
Regards,Les H
Les wrote:
I read the blog, and it said that Fedora Core was going to merge with Fedora Extras (which is the way the DVD is packaged now anyway) and become just Fedora. In other words, the apparent plan is to put to gether the OS, and a suite of tools to make the users computer useful. I think this will make Fedora more attractive, especially if they get the stuff all bound together so that all the chasing down of plugins, dependencies and packages can be done prior to the new package being shipped. Also this should permit tighter control over the entire suite of system and tools, enabling a distribution that will be more or less autonomous and correctly configured at the outset.
Won't this just force the replacement of fedora extras with some other place that is not so tightly controlled?
On 21Dec2006 08:08, les mikesell lesmikesell@gmail.com wrote: | Les wrote: | >[...]I think this will make Fedora more attractive, especially if they get | >the stuff all bound together so that all the chasing down of plugins, | >dependencies and packages can be done prior to the new package being | >shipped. Also this should permit tighter control over the entire suite | >of system and tools, enabling a distribution that will be more or less | >autonomous and correctly configured at the outset. | | Won't this just force the replacement of fedora extras with some | other place that is not so tightly controlled?
Livna, freshrpms, rpmforge and atrpms don't already fill that niche?
"More controlled" looks, to me, like it means "without the horrible dependancy breakages" - they're talking about better consistency checking are the control described. By giving leaf packages more directly to maintainers the "update" flipside of that should become "less controlled", but also less flakey. And it looks like the non-leaf packages will farm out more directly too, later.
On Thu, Dec 21, 2006 at 02:38:03AM -0800, Les wrote:
This will displace windows quite easily especially due to the hassels of getting windows, the support costs, and the licensing fees and so forth that cause windows to be one of the most hated OS's ever produced.
Advocatus Diaboli time. I really don't think Fedora is a Windows- displacement option either at home or in business; and it's not any of the technical issues. Simply put, neither home nor business users can deal with a six-month replacement cycle that obsoletes their base software in a year or so.
We're not just talking about "there's something newer"; we're talking about "you stop getting updates and patches." Fedora Legacy mitigated that; without legacy updates that at least offer the option of not having to do forced updates every year, I think you're optimistic to see Fedora as a Windows-killer.
I had no problems with my last upgrade--but that was one laptop. Look at the problems reported on the list with every new release of Fedora--and imagine you're in IT, and have to support maybe 4-5 variants of workstations, not to mention servers (if you decide to try Fedora there); and not just one of each, but maybe dozens, hundreds, or even thousands.
Or you're a home user, with one to several machines--but you're NOT a tech, just want the computer to balance the checkbook and browse the Internet, with one for your spouse and a kid's homework machine. Face it-- most such users get the OS with the computer, and throw out the computer and OS when it's time to change. Few home users who were stuck with Windows ME or Windows XP Home ever have upgraded (and they usually find-- especially the former--that the hardware can't hack the upgrade.) So if the OS isn't going to remain patched and stable for the 2-3 years or more these people keep their machine, it ain't gonna fly.
So Fedora isn't the Linux Windows Killer. And maybe that's not bad. It is a place for the avant-garde to test the edge; problems are accepted as part of the process, and there is community support and individual efforts to resolve issues.
$0.02, YMMV. -- Dave Ihnat President, DMINET Consulting, Inc. dihnat@dminet.com
Dave Ihnat <dihnat <at> dminet.com> writes:
with Windows ME or Windows XP Home ever have upgraded (and they usually find-- especially the former--that the hardware can't hack the upgrade.)
This is not that big an issue with Fedora. The most memory-hungry part of Fedora is actually the installer, and you can bypass that by using a depsolver like yum or apt on a running system with lots of swap space for your upgrades instead. The CPU is mostly irrelevant for the core OS. (Some apps like OpenOffice.org 2 and Eclipse are another story. But there are more lightweight alternatives, e.g. AbiWord or KWord instead of OO.o Writer. And even here, RAM is more of a limiting factor than CPU power, the main problem is HDD thrashing caused by swapping panic.)
Case in point, I'm running FC6 on a 266 MHz laptop which originally shipped with Window$ 98 (The first exemplaries sold even came with Window$ 95!), with RAM upgraded from 32 MB to the maximum supported (160 MB). I had to fight some interesting battles with Anaconda, so for FC5->FC6 I just used apt-get dist-upgrade and that worked fine (unlike the horror stories from some Ubuntu users). But once installed, KDE just works. I have a more powerful desktop for everyday use, but when I need a laptop, the 266 MHz one works fine.
Kevin Kofler
On Thursday 21 December 2006 17:45, Kevin Kofler wrote:
Dave Ihnat <dihnat <at> dminet.com> writes:
with Windows ME or Windows XP Home ever have upgraded (and they usually find-- especially the former--that the hardware can't hack the upgrade.)
This is not that big an issue with Fedora. The most memory-hungry part of Fedora is actually the installer, and you can bypass that by using a depsolver like yum or apt on a running system with lots of swap space for your upgrades instead. The CPU is mostly irrelevant for the core OS. (Some apps like OpenOffice.org 2 and Eclipse are another story. But there are more lightweight alternatives, e.g. AbiWord or KWord instead of OO.o Writer. And even here, RAM is more of a limiting factor than CPU power, the main problem is HDD thrashing caused by swapping panic.)
Case in point, I'm running FC6 on a 266 MHz laptop which originally shipped with Window$ 98 (The first exemplaries sold even came with Window$ 95!), with RAM upgraded from 32 MB to the maximum supported (160 MB). I had to fight some interesting battles with Anaconda, so for FC5->FC6 I just used apt-get dist-upgrade and that worked fine (unlike the horror stories from some Ubuntu users). But once installed, KDE just works. I have a more powerful desktop for everyday use, but when I need a laptop, the 266 MHz one works fine.
You miss the point, Kevin. Of course you can upgrade on older systems - if you want to. I've done it myself. I have two boxes, though, that run FC4 and for several reasons I do not want to upgrade them. It's just 10 months since FC4 was installed on both of them. I don't need the latest and greatest on either of them, but I do want security updates, and I'm not going to get them. Frankly, Legacy was the biggest reason I had for coming to Fedora. I understand about the lack of manpower in volunteer situations, but I'm less than happy. If I have to install afresh to get a secure system I'll probably change to CentOS rather than install FC6 on those boxes.
Anne
Anne Wilson wrote:
You miss the point, Kevin. Of course you can upgrade on older systems
- if you want to. I've done it myself. I have two boxes, though, that
run FC4 and for several reasons I do not want to upgrade them. It's just 10 months since FC4 was installed on both of them. I don't need the latest and greatest on either of them, but I do want security updates, and I'm not going to get them. Frankly, Legacy was the biggest reason I had for coming to Fedora. I understand about the lack of manpower in volunteer situations, but I'm less than happy. If I have to install afresh to get a secure system I'll probably change to CentOS rather than install FC6 on those boxes.
For what it's worth, there is a current official proposal that security support should be extended to about thirteen months -- support for installing one release, missing the next completely, and then when the release *two* after you'd originally installed came out, you'd have a month to upgrade.
See http://fedoraproject.org/wiki/FedoraSummit/ReleaseProcess (at the bottom in bold) and https://www.redhat.com/archives/fedora-advisory-board/2006-December/msg00099... and the following thread.
Hope this helps,
James.
On Thursday 21 December 2006 20:36, James Wilkinson wrote:
Anne Wilson wrote:
You miss the point, Kevin. Of course you can upgrade on older systems
- if you want to. I've done it myself. I have two boxes, though, that
run FC4 and for several reasons I do not want to upgrade them. It's just 10 months since FC4 was installed on both of them. I don't need the latest and greatest on either of them, but I do want security updates, and I'm not going to get them. Frankly, Legacy was the biggest reason I had for coming to Fedora. I understand about the lack of manpower in volunteer situations, but I'm less than happy. If I have to install afresh to get a secure system I'll probably change to CentOS rather than install FC6 on those boxes.
For what it's worth, there is a current official proposal that security support should be extended to about thirteen months -- support for installing one release, missing the next completely, and then when the release *two* after you'd originally installed came out, you'd have a month to upgrade.
See http://fedoraproject.org/wiki/FedoraSummit/ReleaseProcess (at the bottom in bold) and https://www.redhat.com/archives/fedora-advisory-board/2006-December/msg0009 9.html and the following thread.
I like Rahul's proposal. 13 months is just not quite good enough, because it takes a few weeks after release to get everything settled down. That could leave a server in a less than satisfactory situation for, say, a month. Doing it Rahul's way, servers could be updated annually, a few weeks after release time.
Anne
I've been following this thread for some time and would like to add a few comments. I've been a user of RHL since about version 4 and Slackware since kernel 0.99pl13. I've seen a lot of changes, mostly for the better and really like FC, however I have some concerns (much as I did when RHL -> FC).
We started using FC for several reasons including that it is the extension of RHL, is close to RHEL, is supported by the hardware/software stack we use, is generally stable and secure, had medium-term support via fedora-legacy, and was free. Our current tested and approved baseline is FC4. I am concerned over the loss of fedora-legacy and shortened support.... Maybe we should have used CentOS instead.
The lack of long-term support will hurt Fedora. Why should one install the latest on each and every computer as opposed to just on a few and upgrade every year or so. For bleeding-edge developers, 6 months is OK, but for my wife's computer or my development network 13 months even is way too short. A couple of years would be more reasonable especially if someone is running it on dozens of computers as I am (without sysadmin support).
Last month, I upgraded three computers at home to FC6 from FC5 and FC4 with many issues (posted to the mailing list). One FC5->FC6 upgrade (laptop, x86_64) took nearly 20 hours! Many said that I should just have done a fresh install, but on multiple computers at home, a development network at work, and some machines across the country, that would be difficult. My development network of dozens of computers is mostly baselined on FC4 with a few FC5 test machines. I will not have time until February to begin using FC6 on the development network, yet no updates for FC4. That is reality.
If the update process was fixed or streamlined, it would not be as much of an issue, but 20 hours for 1 computer is a bit too much (or even the 3-4 hours for the other ones I upgraded).
These are some of my Fedora recommendations: 1) streamline the update process so that it does not take much more time than a fresh install - this should encourage updates and would help with adoption 2) fix the many cases where yum update fails due to dependency mess - Fedora will NEVER replace windows if updates require you to manually remove stuff to make it work - I hear complaints about RPM that sound like Windows DLL Hell complaints from the late 1990's! - merge of extras and core should really help here 3) fix the long-standing RPM issue of hanging if you cancel an update or install (__db* files remaining) - very old issue, still an issue with FC6 AFIK, impacts updates 4) support two previous versions for at least 18 months (2 years would be optimal) - For example, I would only have to update the wife's computer every year 5) reduce the requirements for the installer (memory, etc.) for legacy hardware 6) reduce the number of required CDs for a very basic, minimal install to 1 or 2 7) reduce the minimal install footprint (remember the RULE project?) 8) work with mondo archive or similar on a suite of replication and backup capabilities and bundle with FC
Hope my comments help.... -- Wade Hampton
I keep wanting to ask this question, but I'm not quite sure how to phrase it, so here goes...
What is the purpose of releasing different versions of Fedora (Core, or whatever it is called going forward)?
Perhaps a different way to state the same question is: What do I get by upgrading from FC5 to FC6 (to F7, F8, etc..) that I don't get by simply running "yum update"?
Or, trying the question one more way: When I run "yum update", as I do semi-sporadically, why do many packages on my FC5 box get updated so regularly -- they can't all be security updates (can they?)
I'm going to go ahead and send this out like this -- each time I try to write an email with these questions, I get to this point and say to myself, "Well, that's not really want I'm trying to ask", and I delete my draft, get distracted by something else, and am left wondering what the answers are. So, now I'll just let this fly and see what I learn.
--wpd
Patrick Doyle wrote:
I keep wanting to ask this question, but I'm not quite sure how to phrase it, so here goes...
What is the purpose of releasing different versions of Fedora (Core, or whatever it is called going forward)?
When a developer imports a shiny new program into Fedora, or a shiny new version of an existing program, that program is going to have bugs. Sometimes it means that users will have to do things different ways. The developer can find many obvious bugs for himself, but other people are going to use the package in other ways on equipment he doesn't own, and they are likely to find other bugs.
So there's the Fedora development tree, which people use on computers when they don't mind too much if they have to reinstall every so often. They use the computers much as they normally would, or run test-suites, or do whatever, find bugs, and report them. And if they don't like new features, they at least get to say so.
For example, the current development tree contains a kernel which handles parallel IDE drives the same way as serial ATA (SATA) is handled. This means that what was /dev/hda is now /dev/sda (or sdb, or whatever). This will break a number of setups, and mean that some people have to change their configuration files (usually the files they changed themselves). But it does promise better support in the future. Dave Jones, the maintainer, says he won't release this version for FC6 (or FC5), because it would break too many existing setups. But for FC7, people will expect to have to reconfigure a few things, and there may be support in the FC installer for "obvious" settings to be transferred to the new naming.
Perhaps a different way to state the same question is: What do I get by upgrading from FC5 to FC6 (to F7, F8, etc..) that I don't get by simply running "yum update"?
Well, you'll get very few of the new programs doing things in different ways, and very often you'll be limited as to which new programs are available. You won't get major updates.
Or, trying the question one more way: When I run "yum update", as I do semi-sporadically, why do many packages on my FC5 box get updated so regularly -- they can't all be security updates (can they?)
Some are, some aren't. If a new version of a program comes out, and it's not a major update from what was shipped, and it fixes non-security bugs or adds new features, and it's not going to affect a number of other programs, then it will probably be updated during a Fedora release. If it involves rewriting configuration files, or dumping and reloading databases, or changing the way things work, or doing a mass rebuild of a lot of other files, then the Fedora developers will probably put it into the development tree for the next version of Fedora (or the one after that, if there's a version about to be released).
In general, "major updates are for new releases".
Hope this helps,
James.
Or, trying the question one more way: When I run "yum update", as I do semi-sporadically, why do many packages on my FC5 box get updated so regularly -- they can't all be security updates (can they?)
Some are, some aren't. If a new version of a program comes out, and it's not a major update from what was shipped, and it fixes non-security bugs or adds new features, and it's not going to affect a number of other programs, then it will probably be updated during a Fedora release. If it involves rewriting configuration files, or dumping and reloading databases, or changing the way things work, or doing a mass rebuild of a lot of other files, then the Fedora developers will probably put it into the development tree for the next version of Fedora (or the one after that, if there's a version about to be released).
In general, "major updates are for new releases".
Hope this helps,
Thanks. It certainly clarifies a lot... and it broadens the current topic of discussion to include "what is it that people _expect_ out of a Fedora release?" The folks who are saying "Gee, maybe I should switch to CentOS" should probably do that, given Fedora's stated goal of a new major release every 6 months, and given the maintenance requirements of supporting older releases.
--wpd
James Wilkinson wrote:
For example, the current development tree contains a kernel which handles parallel IDE drives the same way as serial ATA (SATA) is handled. This means that what was /dev/hda is now /dev/sda (or sdb, or whatever). This will break a number of setups, and mean that some people have to change their configuration files (usually the files they changed themselves). But it does promise better support in the future. Dave Jones, the maintainer, says he won't release this version for FC6 (or FC5), because it would break too many existing setups. But for FC7, people will expect to have to reconfigure a few things, and there may be support in the FC installer for "obvious" settings to be transferred to the new naming.
Well, you could always use the old kernel if the new one doesn't work. (Quite a lot of distribution kernels don't work for me anyway, on one or other of my machines.)
I think the OP asked quite a pertinent question, and I don't think anyone has given a very cogent answer.
On Fri, 2006-12-22 at 16:52 +0000, Timothy Murphy wrote:
Well, you could always use the old kernel if the new one doesn't work. (Quite a lot of distribution kernels don't work for me anyway, on one or other of my machines.)
I think the OP asked quite a pertinent question, and I don't think anyone has given a very cogent answer.
A more succinct question: why should you _ever_ replace a currently working kernel on existing hardware with anything that is not perfectly backwards compatible (i.e. more than bugfixes)?
I'd like to stay current with firefox, evolution, and a couple of other things but don't understand the need to reinstall a kernel and new mostly untested device drivers just to get them. In the gazillion kernel revs that fedora has rolled out, I can't recall seeing one that made a noticeable improvement in anything that previously worked correctly and have had fairly bad luck with them breaking things that had worked before.
When I get new hardware I don't mind fiddling with an up to date kernel with current drivers. I just don't see the point in having to do that every few months to keep a browser, word processor and email program up to date on an existing system. The point of emulating the stable unix API in the first place should be to isolate the kernel and standard libraries from application developement, so why can't we have distributions that keep them separate? Currently the distributions that try to keep the kernel stable don't keep the applications up to day and the ones that update everything break the kernel and drivers all the time.
I agree with Les on that opinion. I have several machines running redhat 9. There isn't anything earth shattering that I want to do with those machines, and the bug fixes that exist tighten the machine pretty tight. I've never had any problems with those machines, and they've been running for years without even a reboot.
- Donald Tripp dtripp@hawaii.edu ---------------------------------------------- HPC Systems Administrator High Performance Computing Center University of Hawai'i at Hilo 200 W. Kawili Street Hilo, Hawaii 96720 http://www.hpc.uhh.hawaii.edu
On Dec 22, 2006, at 8:54 AM, Les Mikesell wrote:
On Fri, 2006-12-22 at 16:52 +0000, Timothy Murphy wrote:
Well, you could always use the old kernel if the new one doesn't work. (Quite a lot of distribution kernels don't work for me anyway, on one or other of my machines.)
I think the OP asked quite a pertinent question, and I don't think anyone has given a very cogent answer.
A more succinct question: why should you _ever_ replace a currently working kernel on existing hardware with anything that is not perfectly backwards compatible (i.e. more than bugfixes)?
I'd like to stay current with firefox, evolution, and a couple of other things but don't understand the need to reinstall a kernel and new mostly untested device drivers just to get them. In the gazillion kernel revs that fedora has rolled out, I can't recall seeing one that made a noticeable improvement in anything that previously worked correctly and have had fairly bad luck with them breaking things that had worked before.
When I get new hardware I don't mind fiddling with an up to date kernel with current drivers. I just don't see the point in having to do that every few months to keep a browser, word processor and email program up to date on an existing system. The point of emulating the stable unix API in the first place should be to isolate the kernel and standard libraries from application developement, so why can't we have distributions that keep them separate? Currently the distributions that try to keep the kernel stable don't keep the applications up to day and the ones that update everything break the kernel and drivers all the time.
-- Les Mikesell lesmikesell@gmail.com
-- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
Patrick Doyle wrote:
Perhaps a different way to state the same question is: What do I get by upgrading from FC5 to FC6 (to F7, F8, etc..) that I don't get by simply running "yum update"?
Basically, AFAIU, you get major version upgrades. For example, FC5 has GNOME 2.14 as the main Desktop. FC6 has 2.16. FC5 is not going to get 2.16. Ever. It will only get updates for minor versions.
On Fri, Dec 22, 2006 at 10:32:16AM -0500, Dmitriy Kropivnitskiy wrote:
Basically, AFAIU, you get major version upgrades. For example, FC5 has GNOME 2.14 as the main Desktop. FC6 has 2.16. FC5 is not going to get 2.16. Ever. It will only get updates for minor versions.
I think what he's getting at is, why do big-bang releases instead of simply continually releasing updates via an automatic mechanism such as yum?
This model has had its proponents over the years. Probably the biggest reasons you have major releases are:
o A major release gives someone new to the product line a starting point that isn't horrendously out of date. Ever have to reinstall a copy of Windows XP from CD, then live through hours of updates?
o Some changes to the system are so major, affecting underlying functionality, structure, etc., that they're simply too difficult or complex to reliably handle as a dynamic update.
o Having major releases allows the update process to sunset support for old versions of software. Over the years, having to detect and handle every single version of a package that was ever released to properly handle updates would become a nightmare. Add in dependency detection, and, well...
There are other reasons, but these are the biggies.
Cheers, -- Dave Ihnat President, DMINET Consulting, Inc. dihnat@dminet.com
Dave Ihnat wrote:
On Fri, Dec 22, 2006 at 10:32:16AM -0500, Dmitriy Kropivnitskiy wrote:
Basically, AFAIU, you get major version upgrades. For example, FC5 has GNOME 2.14 as the main Desktop. FC6 has 2.16. FC5 is not going to get 2.16. Ever. It will only get updates for minor versions.
I think what he's getting at is, why do big-bang releases instead of simply continually releasing updates via an automatic mechanism such as yum?
This model has had its proponents over the years. Probably the biggest reasons you have major releases are:
o A major release gives someone new to the product line a starting point that isn't horrendously out of date. Ever have to reinstall a copy of Windows XP from CD, then live through hours of updates?
In my experience though, you *do* have to sit through hours of updates: the first few months of a release are so hectic with bug fixes that you end up downloading nearly an entire CD's worth of information in updates to what you just installed. I can do this because I have DSL that can run for hours downloading everything, but it won't work for anyone on dialup.
That brings me to a related question I've wondered for some time: why do we have to download entire packages for updates? Why can't there be an RPM package similar to patches? Then you'd only have to download the difference in a package (and I don't mean a partial file, but just whole files that have been updates. Most files don't get too large individually). I would see where this could be a problem if we didn't have new Fedora Core releases, but since we do, the patch RPMs would only have to be based off the initial package of that version release. If any patch RPMs are needed before a particular patch RPM could be installed, I don't really see why it would be a problem for the patch RPM's spec file to include a list of dependency patches (much like packages already do) and have yum automatically download them too.
Justin W
[snip]
Cheers,
Dave Ihnat President, DMINET Consulting, Inc. dihnat@dminet.com
Quoting Justin W jlist@jdjlab.com:
Dave Ihnat wrote:
On Fri, Dec 22, 2006 at 10:32:16AM -0500, Dmitriy Kropivnitskiy wrote:
Basically, AFAIU, you get major version upgrades. For example, FC5 has GNOME 2.14 as the main Desktop. FC6 has 2.16. FC5 is not going to get 2.16. Ever. It will only get updates for minor versions.
I think what he's getting at is, why do big-bang releases instead of
simply
continually releasing updates via an automatic mechanism such as yum?
This model has had its proponents over the years. Probably the biggest reasons you have major releases are:
o A major release gives someone new to the product line a starting point that isn't horrendously out of date. Ever have to reinstall a copy of Windows XP from CD, then live through hours of updates?
well why wouldn't a respin take care of that? granted a big download would be needed, user intervention would be infrequent, easing the job. a single download, even if large is easier under common circumstances tha
In my experience though, you *do* have to sit through hours of updates: the first few months of a release are so hectic with bug fixes that you end up downloading nearly an entire CD's worth of information in updates to what you just installed. I can do this because I have DSL that can run for hours downloading everything, but it won't work for anyone on dialup.
That brings me to a related question I've wondered for some time: why do we have to download entire packages for updates? Why can't there be an RPM package similar to patches? Then you'd only have to download the difference in a package (and I don't mean a partial file, but just whole files that have been updates. Most files don't get too large individually). I would see where this could be a problem if we didn't have new Fedora Core releases, but since we do, the patch RPMs would only have to be based off the initial package of that version release. If any patch RPMs are needed before a particular patch RPM could be installed, I don't really see why it would be a problem for the patch RPM's spec file to include a list of dependency patches (much like packages already do) and have yum automatically download them too.
Justin W
[snip]
o Some changes to the system are so major, affecting underlying functionality, structure, etc., that they're simply too difficult or complex to reliably handle as a dynamic update.
ok a respin would deal with that too
o Having major releases allows the update process to sunset support for old versions of software. Over the years, having to detect and handle every single version of a package that was ever released to properly handle updates would become a nightmare. Add in dependency detection, and, well...
but isn't sunsetting of sw the issue? "over the years" is currently "over the months", no? If a known and reasonable lifetime (three years for example) is not manageable maybe we ought to look at valuing reliability as a higher priority in design as opposed to fast coding and user-borne QA. We don't need the MS model here...
Dave
Cheers,
Dave Ihnat President, DMINET Consulting, Inc. dihnat@dminet.com
-- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
Dave Stevens wrote: [snip]
o Some changes to the system are so major, affecting underlying functionality, structure, etc., that they're simply too difficult or complex to reliably handle as a dynamic update.
ok a respin would deal with that too
This is essentially what a Fedora Core release is right now.
Rawhide gets the newest packages without caring about disruptions. At fixed point in time a "snapshot" of rawhide is taken and it becomes Test1. Developers and testers work with Test1, Test2, and Test3 to get the disruptions cleared up and make sure everything works together. Then that set of packages becomes the "release".
After the release, rawhide goes back to getting the newest packages without caring about disruptions while the release gets security and bug fixes without disruptive changes.
On Fri, 2006-12-22 at 14:14 -0500, William Hooper wrote:
After the release, rawhide goes back to getting the newest packages without caring about disruptions while the release gets security and bug fixes without disruptive changes.
I think there is a difference of opinion here about how disruptive those changes are... The mailing list has been pretty calm recently but I think every prior release has had outbursts of complaints about previously working hardware breaking from core updates. I realize that these changes have to be tested somewhere and that fedora may be the right place for it to happen, but it is a mistake to think that the distribution is stable enough to depend on for critical work on a large assortment of hardware.
On Friday 22 December 2006 12:11, Justin W wrote:
Dave Ihnat wrote:
On Fri, Dec 22, 2006 at 10:32:16AM -0500, Dmitriy Kropivnitskiy wrote:
Basically, AFAIU, you get major version upgrades. For example, FC5 has GNOME 2.14 as the main Desktop. FC6 has 2.16. FC5 is not going to get 2.16. Ever. It will only get updates for minor versions.
I think what he's getting at is, why do big-bang releases instead of simply continually releasing updates via an automatic mechanism such as yum?
This model has had its proponents over the years. Probably the biggest reasons you have major releases are:
o A major release gives someone new to the product line a starting point that isn't horrendously out of date. Ever have to reinstall a copy of Windows XP from CD, then live through hours of updates?
In my experience though, you *do* have to sit through hours of updates: the first few months of a release are so hectic with bug fixes that you end up downloading nearly an entire CD's worth of information in updates to what you just installed. I can do this because I have DSL that can run for hours downloading everything, but it won't work for anyone on dialup.
That brings me to a related question I've wondered for some time: why do we have to download entire packages for updates? Why can't there be an RPM package similar to patches? Then you'd only have to download the difference in a package (and I don't mean a partial file, but just whole files that have been updates. Most files don't get too large individually). I would see where this could be a problem if we didn't have new Fedora Core releases, but since we do, the patch RPMs would only have to be based off the initial package of that version release. If any patch RPMs are needed before a particular patch RPM could be installed, I don't really see why it would be a problem for the patch RPM's spec file to include a list of dependency patches (much like packages already do) and have yum automatically download them too.
Justin W
I'd like to second those thoughts. OTOH, if one wants to, I believe the basic rsync algorithm could be trained to do exactly that. The disadvantage to that is that one would have to maintain a local rpm repository of the whole install in order to make that work. But a scenario where it sucks a list of files that have been updated since the last time the update was executed, with both the old version and the new version, and add a bit of code to edit the filename in the repo to match, possibly just softlinking the new name to the old file, then just run through it, grabbing what it needs. I could foresee that the network traffic load might be cut by at least 50%, and possibly as much as 95+% for simple patch a function upgrades. The user then would simply run another script that detected the filename update and update-installed the updated package from the users own repository.
This wouldn't take a lot of 'new' code, and could probably be done by a python programmer to be fancy, or even a bash script could probably do it, run at some low load time by cron. The possibilities seem to be well worth the effort IMO.
Dave Ihnat President, DMINET Consulting, Inc. dihnat@dminet.com
Thanks for triggering the thought Dave, & Merry Christmas.
Justin W wrote:
That brings me to a related question I've wondered for some time: why do we have to download entire packages for updates? Why can't there be an RPM package similar to patches? Then you'd only have to download the difference in a package (and I don't mean a partial file, but just whole files that have been updates.
SuSE has something like this, called patch RPMs or delta RPMs: see ftp://ftp.suse.com/pub/projects/deltarpm/README. Fedora has steered away from the concept to date.
Part of the problem with delta RPMs is that they can take up more space and/or more download time: if most of the files have been recompiled against a different version of a library or with a different version of gcc, then one patch RPM could be more than half the size of the full package. If the package is updated again, then you've either got to apply two deltas to get from the original to the newest version (which will be larger than just getting the full RPM), or mirrors[1] have to carry two deltas (original to newest, and update to newest). And the mirror would have to carry the full updated version as well (for people who suddenly want to install the package for the first time), and probably the original package.
Alternatively, you could get the mirrors to generate patch RPMs on the fly. But that requires that they run special software just for the Fedora mirror.
So far, it's been understood that many mirrors would be reluctant either to see the storage requirement for Fedora increase too drastically, or to run custom software just for Fedora.
See also the thread from http://www.redhat.com/archives/rhl-devel-list/2005-March/msg00881.html and http://www.redhat.com/archives/rhl-devel-list/2005-March/msg01031.html.
One thing that *has* happened is that Fedora has moved to more, smaller, more independent packages for certain large pieces of software (OpenOffice, X.org), which means that most updates don't involve storing and downloading the whole kit and caboodle just because one file has changed.
Hope this helps,
James.
[1] Places like ftp.kernel.org, which mirror a copy of the Fedora FTP site, providing bandwidth and disk space.
James Wilkinson wrote:
Justin W wrote:
That brings me to a related question I've wondered for some time: why do we have to download entire packages for updates? Why can't there be an RPM package similar to patches? Then you'd only have to download the difference in a package (and I don't mean a partial file, but just whole files that have been updates.
SuSE has something like this, called patch RPMs or delta RPMs: see ftp://ftp.suse.com/pub/projects/deltarpm/README. Fedora has steered away from the concept to date.
Part of the problem with delta RPMs is that they can take up more space and/or more download time: if most of the files have been recompiled against a different version of a library or with a different version of gcc, then one patch RPM could be more than half the size of the full package. If the package is updated again, then you've either got to apply two deltas to get from the original to the newest version (which will be larger than just getting the full RPM), or mirrors[1] have to carry two deltas (original to newest, and update to newest). And the mirror would have to carry the full updated version as well (for people who suddenly want to install the package for the first time), and probably the original package.
I'm not trying to start a heated debate (just curious :-), but why would you have to have a patch rpm against both the original package and an updated package? Why couldn't it be set up like patches where you would require previous patch rpms to be installed, and we could even obsolete a patch rpm with a later patch rpm, bringing the download total back down again (this makes me think of what SP2 does for XP, eliminating the need for a googolplex KB updates, though I'm not considering the overall size here). I do see the point in an initial install after the release date would require downloading the package and all of the patches, so that could get to be a problem.
[snip]
Justin W
I wrote:
Part of the problem with delta RPMs is that they can take up more space and/or more download time: if most of the files have been recompiled against a different version of a library or with a different version of gcc, then one patch RPM could be more than half the size of the full package. If the package is updated again, then you've either got to apply two deltas to get from the original to the newest version (which will be larger than just getting the full RPM), or mirrors[1] have to carry two deltas (original to newest, and update to newest). And the mirror would have to carry the full updated version as well (for people who suddenly want to install the package for the first time), and probably the original package.
Justin W wrote:
I'm not trying to start a heated debate (just curious :-), but why would you have to have a patch rpm against both the original package and an updated package? Why couldn't it be set up like patches where you would require previous patch rpms to be installed, and we could even obsolete a patch rpm with a later patch rpm, bringing the download total back down again (this makes me think of what SP2 does for XP, eliminating the need for a googolplex KB updates, though I'm not considering the overall size here). I do see the point in an initial install after the release date would require downloading the package and all of the patches, so that could get to be a problem.
I think you've just answered your own question.
Computers will be in any state (as far as a particular package goes): uninstalled, unpatched, or any number of previous patches installed. *Either* you have minimal (possibly delta) RPMs to get users from any of those states to the current updated package, which means one RPM per state, *or* you require some of them to download multiple patch RPMs, which will be larger than the minimal RPM they'd ideally like.
And at some point, the series of RPMs is larger and more of a problem than just having full package RPMs.
James.
On Fri, 2006-12-22 at 08:51 -0500, Wade Hampton wrote:
I've been following this thread for some time and would like to add a few comments. I've been a user of RHL since about version 4 and Slackware since kernel 0.99pl13. I've seen a lot of changes, mostly for the better and really like FC, however I have some concerns (much as I did when RHL -> FC).
We started using FC for several reasons including that it is the extension of RHL, is close to RHEL, is supported by the hardware/software stack we use, is generally stable and secure, had medium-term support via fedora-legacy, and was free. Our current tested and approved baseline is FC4. I am concerned over the loss of fedora-legacy and shortened support.... Maybe we should have used CentOS instead.
The lack of long-term support will hurt Fedora. Why should one install the latest on each and every computer as opposed to just on a few and upgrade every year or so. For bleeding-edge developers, 6 months is OK, but for my wife's computer or my development network 13 months even is way too short. A couple of years would be more reasonable especially if someone is running it on dozens of computers as I am (without sysadmin support).
Last month, I upgraded three computers at home to FC6 from FC5 and FC4 with many issues (posted to the mailing list). One FC5->FC6 upgrade (laptop, x86_64) took nearly 20 hours! Many said that I should just have done a fresh install, but on multiple computers at home, a development network at work, and some machines across the country, that would be difficult. My development network of dozens of computers is mostly baselined on FC4 with a few FC5 test machines. I will not have time until February to begin using FC6 on the development network, yet no updates for FC4. That is reality.
If the update process was fixed or streamlined, it would not be as much of an issue, but 20 hours for 1 computer is a bit too much (or even the 3-4 hours for the other ones I upgraded).
These are some of my Fedora recommendations:
- streamline the update process so that it does not take much more
time than a fresh install - this should encourage updates and would help with adoption 2) fix the many cases where yum update fails due to dependency mess - Fedora will NEVER replace windows if updates require you to manually remove stuff to make it work - I hear complaints about RPM that sound like Windows DLL Hell complaints from the late 1990's! - merge of extras and core should really help here 3) fix the long-standing RPM issue of hanging if you cancel an update or install (__db* files remaining) - very old issue, still an issue with FC6 AFIK, impacts updates 4) support two previous versions for at least 18 months (2 years would be optimal) - For example, I would only have to update the wife's computer every year 5) reduce the requirements for the installer (memory, etc.) for legacy hardware 6) reduce the number of required CDs for a very basic, minimal install to 1 or 2 7) reduce the minimal install footprint (remember the RULE project?) 8) work with mondo archive or similar on a suite of replication and backup capabilities and bundle with FC
---- I thought I would pipe up with a small but necessary commentary here.
People just naturally assume that the goal is to replace Windows. That isn't the goal of Fedora - at least I've never seen that listed as a goal of the Fedora project.
The fact is that Linux in some form(s) will of course replace Windows - http://news.yahoo.com/s/nm/20061214/tc_nm/gartner_prediction_dc
Fedora's objectives are their own and unless/until becoming **the** Windows desktop replacement becomes one of them, all discussion about that is merely one's own projections.
Craig
On Thu, 21 Dec 2006 19:25:44 +0000, Anne Wilson wrote:
Fedora. I understand about the lack of manpower in volunteer situations, but I'm less than happy. If I have to install afresh to get a secure system I'll probably change to CentOS rather than install FC6 on those boxes.
Anne
I agree. I will keep FC for my desktop because I want the latest and greatest and I don't mind rather frequent updates. On the other hand, the servers I manage are still running FC4, and I started a migration plan to go for CentOS a few months ago. Now that Legacy is gone, I will have to accelerate the process.
Akemi
I concur. I use FC for my desktop / laptop, and Scientific Linux (another RHAS rebuild) for our cluster and servers. For high performance computing you are not concerned with the latest greatest version of xwindows or grub, but more concerned with stability. RHAS rebuild projects give us that stability ( I'm still running 2.6.9-34EL ). But for my desktop where I am tweaking this, compile test programs, evaluating software for other users, its good to be on the bleeding edge, even if sometimes you get cut... :-)
- Donald Tripp dtripp@hawaii.edu ---------------------------------------------- HPC Systems Administrator High Performance Computing Center University of Hawai'i at Hilo 200 W. Kawili Street Hilo, Hawaii 96720 http://www.hpc.uhh.hawaii.edu
On Dec 21, 2006, at 10:44 AM, Akemi Yagi wrote:
On Thu, 21 Dec 2006 19:25:44 +0000, Anne Wilson wrote:
Fedora. I understand about the lack of manpower in volunteer situations, but I'm less than happy. If I have to install afresh to get a secure system I'll probably change to CentOS rather than install FC6 on those boxes.
Anne
I agree. I will keep FC for my desktop because I want the latest and greatest and I don't mind rather frequent updates. On the other hand, the servers I manage are still running FC4, and I started a migration plan to go for CentOS a few months ago. Now that Legacy is gone, I will have to accelerate the process.
Akemi
-- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
On Thu, 2006-12-21 at 11:40 -1000, Donald Tripp wrote:
I concur. I use FC for my desktop / laptop, and Scientific Linux (another RHAS rebuild) for our cluster and servers. For high performance computing you are not concerned with the latest greatest version of xwindows or grub, but more concerned with stability. RHAS rebuild projects give us that stability ( I'm still running 2.6.9-34EL ). But for my desktop where I am tweaking this, compile test programs, evaluating software for other users, its good to be on the bleeding edge, even if sometimes you get cut... :-)
I agree with you and your assessment 100%, Donald. The whole deal with Fedora, as noted in the beginning, that it is a hot test bed for RHEL and it will be cyclical, ever evolving at the bleeding edge. No where does the intro to Fedora say that it is a good server candidate.
As I posted before, from what I perceive that I read, FC6 is supposed to be the last edition in that devel cycle and FC7, or whatever they call it, should be settled down and hopefully remain that way for a much longer period of time. It might even be a server candidate, I dunno. I understand that they haven't gone to all the expense to support Fedora for us for our pure bumpin' enjoyment, but maybe Rahul will drop it on us and put all of the speculation to rest. That would be a good thing. Ric
Ric Moore wrote:
As I posted before, from what I perceive that I read, FC6 is supposed to be the last edition in that devel cycle and FC7, or whatever they call it, should be settled down and hopefully remain that way for a much longer period of time.
Ric, I have only been following this thread after the fact, could you give a pointer back to where the above statement was made.
On Mon, 2006-12-25 at 00:31 -0700, Reg Clemens wrote:
Ric Moore wrote:
As I posted before, from what I perceive that I read, FC6 is supposed to be the last edition in that devel cycle and FC7, or whatever they call it, should be settled down and hopefully remain that way for a much longer period of time.
Ric, I have only been following this thread after the fact, could you give a pointer back to where the above statement was made.
http://fedoranews.org/wiki/Fedora_Weekly_News_Issue_67
"FC6 was the final release of Fedora Core. - A much stronger software distribution and community project will take its place. There has never been a better time to get involved in the Fedora Project. Fedora contributors today shape the direction and accelerate the rate of progress of the entire FOSS ecosystem."
Woot! Der it is! Ric
Ric Moore wrote:
I agree with you and your assessment 100%, Donald. The whole deal with Fedora, as noted in the beginning, that it is a hot test bed for RHEL and it will be cyclical, ever evolving at the bleeding edge. No where does the intro to Fedora say that it is a good server candidate.
We have always described Fedora as a general purpose system. Desktop, server or something else - use it the way it fits into your needs.
As I posted before, from what I perceive that I read, FC6 is supposed to be the last edition in that devel cycle and FC7, or whatever they call it, should be settled down and hopefully remain that way for a much longer period of time. It might even be a server candidate, I dunno. I understand that they haven't gone to all the expense to support Fedora for us for our pure bumpin' enjoyment, but maybe Rahul will drop it on us and put all of the speculation to rest. That would be a good thing.
I think you might reading into it more than what is being told. The detailed proposals including IRC logs are in http://fedoraproject.org/wiki/FedoraSummit
Also see the plans for the next release at http://thread.gmane.org/gmane.linux.redhat.fedora.advisory-board/1551
Rahul
Anne Wilson <cannewilson <at> tiscali.co.uk> writes:
I'm less than happy. If I have to install afresh to get a secure system I'll probably change to CentOS rather than install FC6 on those boxes.
Hint: You can force CentOS to upgrade a Fedora installation. "linux upgradeany" is your friend. But the CentOS version has to be significantly newer than the Fedora version for you to get away with this. FC4->CentOS5 will most likely work, FC5->CentOS5 might be problematic due to updates, FC6->CentOS5, FC4->CentOS4 etc. are all recipes for disaster.
Kevin Kofler
Since both distributions use essentially the same packages, just different versions, you can save all your configuration files, re- install, then replace the files with minimal change, if any, and have a CentOS system in a few hours.
- Donald Tripp dtripp@hawaii.edu ---------------------------------------------- HPC Systems Administrator High Performance Computing Center University of Hawai'i at Hilo 200 W. Kawili Street Hilo, Hawaii 96720 http://www.hpc.uhh.hawaii.edu
On Dec 21, 2006, at 10:56 AM, Kevin Kofler wrote:
Anne Wilson <cannewilson <at> tiscali.co.uk> writes:
I'm less than happy. If I have to install afresh to get a secure system I'll probably change to CentOS rather than install FC6 on those boxes.
Hint: You can force CentOS to upgrade a Fedora installation. "linux upgradeany" is your friend. But the CentOS version has to be significantly newer than the Fedora version for you to get away with this. FC4->CentOS5 will most likely work, FC5->CentOS5 might be problematic due to updates, FC6->CentOS5, FC4->CentOS4 etc. are all recipes for disaster.
Kevin Kofler
-- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
On Thursday 21 December 2006 21:53, Donald Tripp wrote:
Since both distributions use essentially the same packages, just different versions, you can save all your configuration files, re- install, then replace the files with minimal change, if any, and have a CentOS system in a few hours.
Thanks, both of you. That's certainly worth consideration
Anne
On Thu, 2006-12-21 at 08:25 -0600, Dave Ihnat wrote:
On Thu, Dec 21, 2006 at 02:38:03AM -0800, Les wrote:
This will displace windows quite easily especially due to the hassels of getting windows, the support costs, and the licensing fees and so forth that cause windows to be one of the most hated OS's ever produced.
Advocatus Diaboli time. I really don't think Fedora is a Windows- displacement option either at home or in business; and it's not any of the technical issues. Simply put, neither home nor business users can deal with a six-month replacement cycle that obsoletes their base software in a year or so.
We're not just talking about "there's something newer"; we're talking about "you stop getting updates and patches." Fedora Legacy mitigated that; without legacy updates that at least offer the option of not having to do forced updates every year, I think you're optimistic to see Fedora as a Windows-killer.
I had no problems with my last upgrade--but that was one laptop. Look at the problems reported on the list with every new release of Fedora--and imagine you're in IT, and have to support maybe 4-5 variants of workstations, not to mention servers (if you decide to try Fedora there); and not just one of each, but maybe dozens, hundreds, or even thousands.
Or you're a home user, with one to several machines--but you're NOT a tech, just want the computer to balance the checkbook and browse the Internet, with one for your spouse and a kid's homework machine. Face it-- most such users get the OS with the computer, and throw out the computer and OS when it's time to change. Few home users who were stuck with Windows ME or Windows XP Home ever have upgraded (and they usually find-- especially the former--that the hardware can't hack the upgrade.) So if the OS isn't going to remain patched and stable for the 2-3 years or more these people keep their machine, it ain't gonna fly.
So Fedora isn't the Linux Windows Killer. And maybe that's not bad. It is a place for the avant-garde to test the edge; problems are accepted as part of the process, and there is community support and individual efforts to resolve issues.
$0.02, YMMV.
I would love Fedora to become a Windows Killer but that won't happen with less than a years support for a distro. I put FC4 on my machines just over a year ago and now I am toast.
Most of those around work here have just done a yum upgrade to FC6 with no issues. I am going to try it next week but expect issues as I have many extra applications installed. Hopefully most have been ported to FC6 or the FC4 ones work.
It would be nice if Fedora is going this route to change the upgrade policy and work as an upgrade instead of suggesting a full install.
Now I did read an article today that because of the changes, there is going to be a team to look at how "rpm" can be made better. Better clean up or working with dependencies would be my first suggestion. Again this is an issue for the SPEC writers from what I have learned to make sure they put the correct info into the file.
I would also like the Fedora team to not mangle the original programs so much. I would like to be able to download the official Fedora application or the real official application without many changes. I look at OpenOffice as an example.
Lets hope that the support for FC5 goes longer than a year.
From: Robin Laing Robin.Laing@drdc-rddc.gc.ca Reply-To: Robin.Laing@drdc-rddc.gc.ca,For users of Fedora fedora-list@redhat.com To: For users of Fedora fedora-list@redhat.com Subject: Re: OS Future now that Fedora Legacy defunct Date: Thu, 21 Dec 2006 14:45:11 -0700
On Thu, 2006-12-21 at 08:25 -0600, Dave Ihnat wrote:
On Thu, Dec 21, 2006 at 02:38:03AM -0800, Les wrote:
This will displace windows quite easily especially due to the hassels of getting windows, the support costs, and the licensing fees and so forth that cause windows to be one of the most hated OS's ever
produced.
Advocatus Diaboli time. I really don't think Fedora is a Windows- displacement option either at home or in business; and it's not any of the technical issues. Simply put, neither home nor business users can deal with a six-month replacement cycle that obsoletes their base software in a year or so.
We're not just talking about "there's something newer"; we're talking about "you stop getting updates and patches." Fedora Legacy mitigated that; without legacy updates that at least offer the option of not
having
to do forced updates every year, I think you're optimistic to see Fedora as a Windows-killer.
I had no problems with my last upgrade--but that was one laptop. Look at the problems reported on the list with every new release of Fedora--and imagine you're in IT, and have to support maybe 4-5 variants of workstations, not to mention servers (if you decide to try Fedora there); and not just one of each, but maybe dozens, hundreds, or even thousands.
Or you're a home user, with one to several machines--but you're NOT a tech, just want the computer to balance the checkbook and browse the Internet, with one for your spouse and a kid's homework machine. Face it-- most such users get the OS with the computer, and throw out
the
computer and OS when it's time to change. Few home users who were stuck with Windows ME or Windows XP Home ever have upgraded (and they usually find-- especially the former--that the hardware can't hack the upgrade.) So if the OS isn't going to remain patched and stable for the 2-3 years or more these people keep their machine, it ain't gonna fly.
So Fedora isn't the Linux Windows Killer. And maybe that's not bad. It is a place for the avant-garde to test the edge; problems are
accepted
as part of the process, and there is community support and individual efforts to resolve issues.
$0.02, YMMV.
I would love Fedora to become a Windows Killer but that won't happen with less than a years support for a distro. I put FC4 on my machines just over a year ago and now I am toast.
Most of those around work here have just done a yum upgrade to FC6 with no issues. I am going to try it next week but expect issues as I have many extra applications installed. Hopefully most have been ported to FC6 or the FC4 ones work.
It would be nice if Fedora is going this route to change the upgrade policy and work as an upgrade instead of suggesting a full install.
Now I did read an article today that because of the changes, there is going to be a team to look at how "rpm" can be made better. Better clean up or working with dependencies would be my first suggestion. Again this is an issue for the SPEC writers from what I have learned to make sure they put the correct info into the file.
I would also like the Fedora team to not mangle the original programs so much. I would like to be able to download the official Fedora application or the real official application without many changes. I look at OpenOffice as an example.
Lets hope that the support for FC5 goes longer than a year.
-- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
I think they also need to find a way to auto detect settings for applications that are already compiled/installed...and recompile them on the upgrade.
..apache, postfix etc..
Jim
_________________________________________________________________ Get live scores and news about your team: Add the Live.com Football Page www.live.com/?addtemplate=football&icid=T001MSN30A0701
As I posted on a few other times, I did an update (hard disk images) from FC5 to FC6 on my laptop (x86_64) and it took about 15 hours. Two other updates, again from HD images (X86_64 took Updates need to be smoother if the OS is not going to be supported for more than 6-9 months.
On 12/23/06, Jim Douglas jdz99@hotmail.com wrote:
From: Robin Laing Robin.Laing@drdc-rddc.gc.ca Reply-To: Robin.Laing@drdc-rddc.gc.ca,For users of Fedora fedora-list@redhat.com To: For users of Fedora fedora-list@redhat.com Subject: Re: OS Future now that Fedora Legacy defunct Date: Thu, 21 Dec 2006 14:45:11 -0700
On Thu, 2006-12-21 at 08:25 -0600, Dave Ihnat wrote:
On Thu, Dec 21, 2006 at 02:38:03AM -0800, Les wrote:
This will displace windows quite easily especially due to the
hassels
of getting windows, the support costs, and the licensing fees and so forth that cause windows to be one of the most hated OS's ever
produced.
Advocatus Diaboli time. I really don't think Fedora is a Windows- displacement option either at home or in business; and it's not any of the technical issues. Simply put, neither home nor business users can deal with a six-month replacement cycle that obsoletes their base software in a year or so.
We're not just talking about "there's something newer"; we're talking about "you stop getting updates and patches." Fedora Legacy mitigated that; without legacy updates that at least offer the option of not
having
to do forced updates every year, I think you're optimistic to see
Fedora
as a Windows-killer.
I had no problems with my last upgrade--but that was one laptop. Look at the problems reported on the list with every new release of Fedora--and imagine you're in IT, and have to support maybe 4-5 variants of workstations, not to mention servers (if you decide to try Fedora there); and not just one of each, but maybe dozens, hundreds, or even thousands.
Or you're a home user, with one to several machines--but you're NOT a tech, just want the computer to balance the checkbook and browse the Internet, with one for your spouse and a kid's homework machine. Face it-- most such users get the OS with the computer, and throw out
the
computer and OS when it's time to change. Few home users who were
stuck
with Windows ME or Windows XP Home ever have upgraded (and they
usually
find-- especially the former--that the hardware can't hack the
upgrade.)
So if the OS isn't going to remain patched and stable for the 2-3
years
or more these people keep their machine, it ain't gonna fly.
So Fedora isn't the Linux Windows Killer. And maybe that's not bad. It is a place for the avant-garde to test the edge; problems are
accepted
as part of the process, and there is community support and individual efforts to resolve issues.
$0.02, YMMV.
I would love Fedora to become a Windows Killer but that won't happen with less than a years support for a distro. I put FC4 on my machines just over a year ago and now I am toast.
Most of those around work here have just done a yum upgrade to FC6 with no issues. I am going to try it next week but expect issues as I have many extra applications installed. Hopefully most have been ported to FC6 or the FC4 ones work.
It would be nice if Fedora is going this route to change the upgrade policy and work as an upgrade instead of suggesting a full install.
Now I did read an article today that because of the changes, there is going to be a team to look at how "rpm" can be made better. Better clean up or working with dependencies would be my first suggestion. Again this is an issue for the SPEC writers from what I have learned to make sure they put the correct info into the file.
I would also like the Fedora team to not mangle the original programs so much. I would like to be able to download the official Fedora application or the real official application without many changes. I look at OpenOffice as an example.
Lets hope that the support for FC5 goes longer than a year.
-- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
I think they also need to find a way to auto detect settings for applications that are already compiled/installed...and recompile them on the upgrade.
..apache, postfix etc..
Jim
Get live scores and news about your team: Add the Live.com Football Page www.live.com/?addtemplate=football&icid=T001MSN30A0701
-- fedora-list mailing list fedora-list@redhat.com To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
Jim Douglas wrote:
I would also like the Fedora team to not mangle the original programs so much. I would like to be able to download the official Fedora application or the real official application without many changes. I look at OpenOffice as an example.
This *is* something that Fedora are trying to do -- to do as much development as possible "upstream". But there are certain areas (and OpenOffice is *definitely* one of them) where that hasn't been possible, without compromising some of Fedora's principles and other priorities.
OpenOffice, for example, comes from Sun and is increasingly reliant on Sun's Java. This is not Free or Open Source (yet)[1], so that part of the package Just Can't Be Installed without modification.
Then there are things like better support for internationalisation or better integration with the rest of the system, which are very difficult to support across every version of Linux out there -- it's easier for the OpenOffice team to provide their own. But when you know that your OpenOffice will be running on Fedora, you can reasonably make a better package for your end users by tying it into what you know is available on Fedora.
Then there are additions which Red Hat want, and are prepared to fund (e.g. Xen). We should say no? When *all* of Fedora's funding comes from Red Hat (bar some of the infrastructure?)[2]
I think they also need to find a way to auto detect settings for applications that are already compiled/installed...and recompile them on the upgrade.
..apache, postfix etc..
That sounds incredibly weird, incredibly difficult to program, and not what most people want. I mean, why would you *want* to recompile programs? Answers beginning "To support..." miss the point -- ideally, you should be able to get the support by editing a config file or by installing another RPM. Wouldn't it make more sense for Fedora to spend engineer time making that possible?
Fedora *isn't* supposed to support every weird taste. If you want *that* sort of support for locally-compiled programs, you know where Gentoo is!
James.
[1] Anyone saying "but Sun's just released it as GPL" hasn't been paying attention. No. They haven't. They've said they're *going* to, and I believe them, but it Hasn't Happened Yet, and probably won't until it's too late for Fedora 7.
[2] Thanks, guys!
Ric Moore wrote:
On Tue, 2006-12-19 at 21:57 -0600, Justin W wrote:
Shawn wrote:
Once it's released, I'd like to get the equivalent version of CentOS for my server, but if it's going to be a while (say, longer than a month from now),
CentOS wont be ready for at least a month after RHEL5. Has taken 2-3 months in the past.
Shawn
Thank you. I finally found an article which said RHEL 5 should be released about mid-February, and with a 2 or 3 month window before CentOS is released, it means it could be May before I can get a hold of it. That very nicely coincides with summer break. I think I'm going to take the time to install FC6 to my server, and that will also coincide pretty closely with a theoretical end of support for FC6 (assuming 6 month life spans).
You might want to stick with Fedora as the announces also say that Fedora is to become static and meshed with RHEL in some fashion. Ric
Well, at any means, I need to upgrade (I'll put off deciding whether I'm going to go with CentOS or stick with Fedora for another day). I could use some help getting ready for the move.
I've done upgrades before, but they've always been kind of long and I don't like going without my email server for that long. I'm now thinking that I could install FC6 to a larger hard drive that I have acquired and use FC6's Xen capabilities to host the FC4 (from my other HD) while I'm in the process of upgrading. I don't know much about Xen though.
1) Is kernel-xenU the correct kernel I need, or do I have to try to compile one myself?
2) Will Xen run FC4 from the hard drive it's already on, or does it want it as a file in the file system? (FC4 is under LVM if that makes a difference)
3) Will there be any major sticklers I need to worry about when FC4 boots as a guest rather than on its own? Will network connections will be broken in any way? Will I have to worry about my swap partitions (as I know I don't have enough ram to run two systems at once without using swap)?
4) Can somebody help me find a correct xen configuration that will allow my FC4 system to run as it was (i.e. same IP address, etc). I'll read up on some of the xen manuals, using http://www.linuxdevcenter.com/pub/a/linux/2006/01/26/xen.html a reference to see what I need to do.
After I have [ideally] both systems running, I'll slowly move services from one system to the other until I have everything ported and reconfigured (as FC4 was my first server setup, so therefore is pretty messy and needs to be reworked from the ground up).
If anybody wants anything more specific, I'll be happy to give you the information you need.
Thanks, Justin W
Justin W wrote:
Well, at any means, I need to upgrade (I'll put off deciding whether I'm going to go with CentOS or stick with Fedora for another day). I could use some help getting ready for the move.
I've done upgrades before, but they've always been kind of long and I don't like going without my email server for that long. I'm now thinking that I could install FC6 to a larger hard drive that I have acquired and use FC6's Xen capabilities to host the FC4 (from my other HD) while I'm in the process of upgrading. I don't know much about Xen though.
- Is kernel-xenU the correct kernel I need, or do I have to try to
compile one myself?
- Will Xen run FC4 from the hard drive it's already on, or does it
want it as a file in the file system? (FC4 is under LVM if that makes a difference)
- Will there be any major sticklers I need to worry about when FC4
boots as a guest rather than on its own? Will network connections will be broken in any way? Will I have to worry about my swap partitions (as I know I don't have enough ram to run two systems at once without using swap)?
- Can somebody help me find a correct xen configuration that will
allow my FC4 system to run as it was (i.e. same IP address, etc). I'll read up on some of the xen manuals, using http://www.linuxdevcenter.com/pub/a/linux/2006/01/26/xen.html a reference to see what I need to do.
I knew I was forgetting one: 5) How does the boot up process work? How do I have to reference the kernel: as if it were chrooted so just /boot/kernel... or is it /path/to/fc4/root/boot/kernel...?
After I have [ideally] both systems running, I'll slowly move services from one system to the other until I have everything ported and reconfigured (as FC4 was my first server setup, so therefore is pretty messy and needs to be reworked from the ground up).
If anybody wants anything more specific, I'll be happy to give you the information you need.
Thanks, Justin W