Perhaps OpenSuSE needs an 'LTS' version?

One thing that seems to be the clincher for Zimbra dropping
support of OpenSuSE is the perceived lack of commercial support.
Canonical supports Ubuntu and even offers an LTS (Long term support)
version of its desktop and server offerings. Actually it seems rather
like a good idea for an open source offering to have. I see it potentially
NOT being in Novell’s interest to have such a beast floating about, but
if they followed the Canonical model and sold support for the open source
LTS version they could still make a buck on it.

Just curious to hear opinions. I know the SLES/SLED crowd will be up in
arms declaring heresy and that we all need to be forking over for a
real ‘enterprise OS’ but for OpenSuSE as a distro I think it makes sense
and could potentially increase the userbase in the server arena.

Suse Enterprise has long term support.

This is easy for a third party to do, just get a team of people willing to backport security fixes to old unsupported distros, and then provide a repository. Of course easier said than done, and where’s the money going to come from, but still, the infrastructure is already here.

  • GofBorg,

fully agree. IMO the Ubuntu business model is smarter than Novell’s. They didn’t change the Suse approach of a separate business desktop when they bought the company.

Uwe

Hehe, the openSLED story is back again !

My opinion in short : If you want free (as a beer) Long Term Support, use CentOS, or Debian (or UbuntuLTS).

Remember that Ubuntu’s economic model is actually not viable on long term (and maybe never will).

When you accept the concept of openSUSE, the limitations in resources, and
that openSUSE is not highest priority for Novell, then you can start to
discuss how to make it better, realistically within these boundaries.

Quote from our friend Cb400f on Re: [opensuse-project] What things can we do to make openSUSE the most s. It’s not exactly targeted on the cycle life of openSUSE (but on Stability/Innovation), but there is a truth inside these words : Be realistic.

> Hehe, the openSLED story is back again !
>
> My opinion in short : If you want free (as a beer) Long Term Support,
> use CentOS, or Debian (or UbuntuLTS).

I find it ridiculous how many people simply say move to another distro.
You are, in fact, driving users away byt hose types of statements…not
good. I’ve worked with Debian and Ubuntu and IMHO they don’t hold a candle
to OpenSuSE as a server. CentOS, perhaps is an option, but I’d prefer
SuSE.

There doesn’t need to be OpenSLED versus OpenSLES (***note: there’s really
no difference between Desktop and Server distros other than base packages
installed, which are completely customizable at install time anyway), all
there needs to be is an OpenSuSE LTS, which freezes the package level at
some point in time and only supplies critical updates and patches. So you
would have OpenSuSE continue as it is today, with the added LTS benefit for
those who need a longer life cycle. A side effect may be that it increases
commercial support for the OS, another oft mentioned shortcoming of Linux
in general.

> This is easy for a third party to do, just get a team of people willing
> to backport security fixes to old unsupported distros, and then provide
> a repository. Of course easier said than done, and where’s the money
> going to come from, but still, the infrastructure is already here.

Well it’s been mentioned by other before, why does Ubuntu have so
many ‘flavors’ yet OpenSuSE does not? I guess the question applies here as
well. Why isn’t there a third party doing anything with OpenSuSE? I think
it’s because there hasn’t been a need, OpenSuSE has up to this point done
what it needs to, but if it’s long term strategy is simply another desktop
OS, it doesn’t have a very compelling story. IMHO, there needs to be a
server strategy, Ubuntu realizes that and note that they are currently
posing a survey on their site about what do you think should be in a server
OS. They are serious…very serious. Sorry for the Ubuntu comparisons as
I know it’s a hot spot right now. I DO NOT think Ubuntu has the superior
product, but I DO believe they intend to capture enough mindshare to
marginalize other distros, and to that end, their strategy has merit.

That’s all I have to say on the subject…

Off to play with CentOS today.

As I said (or quoted), openSUSE have some limitations in resources.

Sure I would be happy to have a free “openSLE” with LTS, but I don’t want to decrease quality of the current openSUSE.

Backporting security patches during 5 or 7 years is an long, hard, and ingrate work. If someone in the community (not the Novell security staff) want to do it, for free and for all other, just do it.

But please don’t ask weird things, be realistic inside openSUSE boundaries.

PS: Have fun with CentOS, it’s probably actually the best option for a free LTS server.

Aside from a few major flavours like U, K and X, the others are mostly remixes or extra goodies throw in. Some are small experiments that will disappear in time due to lack of interest of the developer, or maybe they might carve a permanent niche. I was just playing with CrashBangLinux (Ubuntu with blackbox WM, nicer than Xubuntu I would say, and smaller footprint). All as it should be. They benefit from a core set of packages, for example there is no need for a separate postfix package for each flavour. Secondly only a few releases are LTS (only 2 so far?), the others are standard 18 months (less than OpenSUSE’s 2 years). So while it is still a major operation and full time job for many developers, it’s not as big as it looks at first.

SuSE came late to the game where remixing is concerned because for a long time SuSE was a box set distro. Only now has it caught up and become also a fully online distribution, upgrade and update distro. One might also argue that SuSE is so good and complete that there is little need for a remix. Where Ubuntu has U, K and X, these GUIs are all standard in SuSE. You can see that there have only been a few remixes of SuSE.

I sympathise with the sentiment that the support lifetime for OpenSUSE should be longer to make it more attractive to be a server platform. OpenSUSE is a good server platform, as so many packages are just already there. But this great variety also makes it difficult to maintain support. Enterprise distros are whittled down to a smaller set of packages. Maybe if the community were to take on LTS of OpenSUSE, only a subset of packages should be supported. But which subset?

On the other hand I think that regular change is a good thing too. It is depressing for me to have to support old servers because I sometimes have to find a way to install new libraries or packages. I should add that some of the servers I support change apps over time, unlike enterprise servers where they deploy them and then don’t touch them afterwards. So while CentOS and RHEL are good for such things, I run into problems with them. Did you know that officially PHP on RHEL5 is still at 5.1? Ditto CentOS of course. Which meant missing addons and in one case even a PHP bug. I had to find some unofficial packages to upgrade to 5.2. Yesterday’s adventure was finding packages for Ruby on Rails. I had to get one SRPM from Fedora and rebuild.

So ded if you do and ded if you don’t. Keep up to date that is. That’s life.

For me i think that suse support(correct me if i’m wrong)of 2 years is long enough.:expressionless:

This is from wikepedia
openSUSE

Main article: openSUSE
openSUSE is a community project, sponsored by Novell, to develop and maintain SUSE Linux components. After their acquisition of SUSE Linux, Novell has decided to make the community an important part of their development process. It is Novell’s goal to market openSUSE as the best, easiest distribution for all users

According to this it looks like opensuse is a test bed for suse enterprise
If this is correct then a LTS would be counterproductive

Yes, I think Novell wouldn’t be too inclined to put money into LTS OpenSUSE. But being open there are no barriers to some other party doing this. You might have to call it something else due to trademarks, like RH and CentOS. But as I said, resourcing is the real problem. Just look at the constant stream of fixes for say 11.0. I’m getting an average of 1 update per day or two, and that’s just the security fixes. I hate to think of the functionality fixes. And they also have to be regression tested to make sure that nothing breaks. It’s not something a bunch of people can do in their spare time. That recent DNS upgrade where every distro had to scramble? LTS OpenSUSE would also have had to scramble too. (Albeit, building a new RPM from an updated SRPM.) I mean if you say LTS, then people will start to rely on it, otherwise what’s the point. And it’s depressing work, as a builder you think why don’t people just update, why am I trying to understand these diffs trying to retrofit a security fix to an old version. But that’s what you have to do if you want a stable release, you cannot change the mix of packages from the original.

In short I have just described what the maintainers of the SLE distros do. And Novell pays their salaries.

So I don’t think LTS OpenSUSE will fly.

My reply to the OP would be an emphatic “Yes, they do”, as their relatively rapid rollouts (pardon the alliteration) of new releases is seriously irritating me, for one…And I know I’m not alone in this, as I’m personally aware of many others in positions in business and Government which have taken issue even with their Business/Enterprise support offerings, and which in two specific cases I can think of, actually lost that business to RHEL solutions…And in one of those cases, we’re talking about well over a thousand desktop versions of RHEL, and a large number of ES/AS/Server versions, in addition.

While nothing in IT is ever really “set it and forget it”, having started out as a ‘Unix’ person some time ago, one of the many reasons I’ve enjoyed the bloom of Open Source, and Linux in particular, was because one could set up a server and have some amount of relief in knowing that – unlike certain cough MS cough Server Operating Systems, there wouldn’t be new remote root exploits every other day, and that, properly configured and hardened – and provided patching was a viable option – one could concern oneself with upgrading only when radically new features or the like were necessary, and otherwise, one had built something which could last for some years, even…Which is nearly ‘forever’, when it comes to IT.

Lest anyone think me uninformed, my first introduction to Linux (beyond the early Usenet postings in 1991 or so) was a friend from NIST giving me a copy of Slackware 3.2, which I played around with at the time, even setting up a test machine, which I found delightfully easy to configure, and which ran very well, even compared to many other things I was evaluating at the time (e.g., SCO et al), although I didn’t deem Linux, at that time, ready for production-level use. I actually began using Slackware at home, and continued following the quickly multiplying distros as they came (and went).

Fast-forward a few years: while Redhat remained my least-favorite distribution, it was the one I found myself dealing with the most in my career, so I was/am reasonably adept with it, and when it came to small/medium-sized businesses, I found the push to RHEL-only (and the EOLing of the previous versions of RH) rather off-putting. This was prior to ‘OpenSuSE’, with ‘SuSE X.x’ being the available items (and the Enterprise-type offerings), and after many trials and evaluations, I was even encouraging the use of SuSE linux to many people I knew, having begun using it myself circa version 8.2, which I ran for some time before migrating that particular workstation to 9.x (eventually 9.3).

Having a number of licenses purchased for 10.x which, because of extenuating circumstances in life, I lacked the time to actually work with/deploy, I now find myself seeing them all EOL’d (or near that point); I made further reflections on whether or not I wished to continue working with this distribution, and find the comments made about going to other distros summarizing my feelings on the topic: perhaps CentOS or Ubuntu’s LTS is the way to go :frowning:

That saddens me, as I’ve grown somewhat fond of SuSE/OpenSuSE, but this MS-like creeping-featurism constant new release business (and subsequent shortening of the span of support for previous versions) isn’t giving me much wiggle room.

How many others feel the same way? How many people at Novell are aware that (not even counting security), one of the many reasons people, organizations, businesses, etcetera would choose a *NIX-based OS for machines is because they DO NOT want to be in a state of constant upgrade?

Just my $0.02 worth.

Regards,

~JMB

I think you have to remember a few things.

CentOS is just rebranded RHEL so if you regard that as suitable for your needs, it’s just the price of RHEL you are not willing to pay. Ubuntu LTS is new to the scene though.

Prior to acquisition by Novell there was no such thing as LTS for S.u.S.E (yes, with that strange punctuation). S.u.S.E also put out new box sets regularly. So Novell hasn’t taken away LTS, it was never there prior to the enterprise distros.

One might have rosy memories of box sets with the numbers 7.1 or 8.2, but those were simpler times also. If you were happy to just host an Apache server with static content, or a Samba file server, then you could go on using the same box and maybe compile from source now and then if some hole is discovered. Unfortunately nowadays developers want their PHP, their AJAX platforms, and so forth. And these are subject to rapid version update. The kernel and the surrounding utilities are fine, they will run indefinitely if the hardware does. I know about this 10.1 box on an internal LAN in a SME. It does fine as a file server and I’m going to leave it at that. But it they start asking to host say Ruby on Rails on it, then I’ll have to point out the stark choices they face: either build the RoR ecosystem from source, or upgrade.

The comparison betwen RHEL and SLE* I’ll leave for others to debate. Certainly SLE* has less of a third party community around it than RHEL, which had the first mover advantage, and has groups like rpmforge. Perhaps one will form around the build service.

These are my personal experiences; these may or may not reflect the views of others.

Ubuntu LTS is overrated. 3 years support on the desktop? You may aswell call OpenSuSE ‘LTS’ as it is. The LTS releases are hardly enterprise-worthy, Dapper for example was a disgrace with a leaky KDE that ate RAM.

As far as RHEL/CentOS goes… RHEL/CentOS 5 has a ton of problems on the desktop. Printing a 50 page document on a 2.0ghz Athlon64 with 1GB RAM on probably THE most universal printer going (HP Deskjet 940C) should work flawlessly. The reality however is a system that slowly leaks and grinds to a halt without printing a single page…

A point in it’s favour over Ubuntu is that it supports the Desktop for 7 years, which is truly long-term.

However, it is very easy to keep a stable stream with OpenSuSE, simply keep one or two versions behind latest (e.g. upgrade from 10.3 -> 11.0 when 11.1 is released).

  • netspionage,

to me it sounds like you are talking about Opensuse. Are you aware of the fact that Novell sells a business version of Opensuse named Suse Linux Enterprise Desktop? The support lifecycle for SLED is much longer, and you get more support.

Uwe

> Ubuntu LTS is overrated. 3 years support on the desktop? You may
> aswell call OpenSuSE ‘LTS’ as it is. The LTS releases are hardly
> enterprise-worthy, Dapper for example was a disgrace with a leaky KDE
> that ate RAM.

There’s nothing I like about Ubuntu, but as it wears an LTS badge, the
reality is that third party vendors are looking at it and saying, Okay here
is a distro that we can depend on it being relatively the same and
supported for 3 years. OpenSuSE while it may be stable for three years
isn’t meeting the criteria that vendors are looking for.

> As far as RHEL/CentOS goes… RHEL/CentOS 5 has a ton of problems on the
> desktop. Printing a 50 page document on a 2.0ghz Athlon64 with 1GB RAM
> on probably THE most universal printer going (HP Deskjet 940C) should
> work flawlessly. The reality however is a system that slowly leaks and
> grinds to a halt without printing a single page…
>
> A point in it’s favour over Ubuntu is that it supports the Desktop for
> 7 years, which is truly long-term.

Yes, and RedHat is widely supported via third parties.

> However, it is very easy to keep a stable stream with OpenSuSE, simply
> keep one or two versions behind latest (e.g. upgrade from 10.3 -> 11.0
> when 11.1 is released).

Well, that is actually how I run my desktops and the OpenSuSE servers I
have. It has served my needs extremely well up until I ran across a
particular vendor whose claims were made that 1) OpenSuSE has no LTS
edition and 2) It is not commercially supported by the Vendor/Backer (aka
Novell) In reality I think only Ubuntu can make that claim. Does RedHat
offer Fedora support? I confess I have no knowledge in that area.
And lastly they claim there was insufficient download interest in the
OpenSuSE version.

FWIW, I have moved on to CentOS due to this issue, but it is running in a VM
on OpenSuSE. :slight_smile:

> to me it sounds like you are talking about Opensuse.
We are on Opensuse.org yes? :slight_smile:

  1. It is not commercially supported by the Vendor/Backer (aka Novell) In reality I think only Ubuntu can make that claim

Just to clear up a misconception. OpenSuSE is supported commercially by Novell as an end-user product. If you buy OpenSuSE you get 90 days technical support covering installation and configuration.

Does RedHat offer Fedora support? I confess I have no knowledge in that area

They do not. That is what sets OpenSuSE and Fedora apart. Fedora is a power user and/or hobbyist product which has only community support while OpenSuSE does have basic end-user technical support with a paid boxed copy.

The Fedora community support is brilliant though, I don’t think i’ve ever had any problems that couldn’t be answered by the guys in their freenode IRC channel.

people who want long support, buy opensuse enterprise

opensuse is a testing bed for that, and should not have long term support.

netspionage, where are you in a constant state of upgrade? You aren’t unless you decide to use openSUSE which is meant as a test bed for Suse Enterprise. And the people who are on OpenSuse typically are because they like being at the edge of the product. Those who are after a more stable product for their usage go for Suse Enterprise.