Oh, God. I just got a dedicated server with it. I see no reason to name the webhosting company here that does this. Anyway, they told me that there is no way to install Ubuntu 8.0 instead of OpenSuSE10 on the server.
Novell Suse Linux Enterprise Server (SLES), current version 10: Novell’s enterprise Linux server product
Novell Suse Linux Desktop (SLED), current version 10: Novell’s enterprise Linux desktop product
Both products require a fee to get maintenance and updates. You get support for a couple of years, certifications from hardware vendors, all the stuff you expect if you are using Linux in a business. Let me know if you need details. Forums are on forums.novell.com
AFAIK Ubuntu has a different business model. The closest in the Ubunty world would be a LTS version plus support.
Then there is Opensuse, the non-commercial community project sponsored by Novell. That’s what we deal with here on forums.opensuse.org. Current version of Opensuse is 11.0.
Does that make things clearer?
The answer to the question on how to update your system depends on which product you use or want to use.
Was it explained to you why you could not upgrade Ubuntu?
There is an exact equivalent in openSUSE to “apt-get upgrade” - it’s “zypper dist-upgrade”. But you will encounter a significant number of issues to resolve, starting with the software package manager being entirely different now. Beyond that, there are changes with the kernel and device handling that are not trivial. Unless you really understand linux and openSUSE, the system will get broken - just as with Ubuntu (or Debian).
If you have been using Ubuntu as a server, perhaps you can plan a migration of the specific applications you’ve been using. If the apps are the same (e.g., apache), you probably can do a clean install and then copy in the control files. Usually newer versions of the software are backward compatible; then you would use the application to update those files. If you are using the system as a mail server however, that migration will be much more challenging and require expertise.
as for reason why Ubuntu can’t be installed–I’ve got no intelligible answer from Webhosting. However smart heads are advising me to use Debootstrap to teleport into Ubuntu. Quite risky, but perhaps a solution…
regarding the rest, my web-application runs over appache and I could manage it under OpenSuSE too. My concern is rather security for the layer that is quite common for different Linux distros that is for apache, psql and of course the whole networking. I used shallowly but long enough SuSE 6-8, RedHat, Fedora, Gentoo to get slowly through the yet another distro.
**Finally, I don’t care what Linux I have, if it is still actively supported by large community **
Well, the questions are:
How to install safely Ubuntu having a Dedicated Server that already runs OpenSuSE 10?
Either one of the alternatives are technically possible. However, whether the choices are feasible or practical are entirely different questions.
Can Ubuntu be installed where openSUSE is already running? Sure, as long as there is adequate storage space and the version of Ubuntu supports the hardware. Debootstrap is simply the bottoms-up hands-on method of installing Debian (the Debian and Ubuntu installers are basically shells atop debootstrap). If you know what you are doing and have a reliable network connection, it’s not particularly difficult. Now, that doesn’t necessarily translate into it being implementable or supportable or a good idea from a business perspective, which may be the concerns of your hosting provider. Ubuntu particularly at the server level is really just Debian, and things are done differently in Debian as compared to Fedora/SuSE/Mandriva (which share the same origins and are architecturally similar). Another concern may actually be security; Fedora and SuSE being upstream for enterprise-class distros (i.e., RH & SLE) includes functions out-of-the-box that may have to be added in by hand in Debian; in any event, while the end result may be the same how one gets there and the expertise required for sustaining support is different. The applications layer (LAMP) will essentially be the same, and an experienced systems administrator can migrate the the configuration and data from one distro to another. So while it is technically possible, but there can be reasons why it is not practical or doable based on circumstances and resources.
The answer to the second question is basically the same. With just a LAMP server, the preferable approach would be a clean install of 11.0 - you want the newest kernel receiving patches, newer security and audit tools, etc. - and then migrate in the applications pieces. In the upgrades discussed on these forums, the context is nearly always desktop users, which is to say, concerns with a plethora of different applications and the whole desktop gui layer, in the hands of users typically with little or no technical experience. But with a LAMP server, the context and concerns are totally different; much fewer but much more complex and typically with more serious (i.e., business or institutional or community) implications. The job must be done right.
In short, the issue is not the software per se; any of these tools can satisfy the technical requirements and the scale of the work is probably comparable. It’s a matter of having available the necessary resources and expertise to do the job, and of course, possibly conditions or requirements that must be met (such as supportability or business cost-effectiveness) which may dictate one choice over another.
I gather from your post that (1) the machine is physically remote, which does add to the challenge, and (2) you personally are providing the necessary expertise I ref’d earlier.
So it comes down to whether you can navigate through the process. If you can handle debootstrap, there is no reason IMO why - with doing the necessary reading and getting community support at Ubuntu or Debian - you can’t do the rest. The amount of work doing the migration, once the basic installation is set up, is not trivial, but it’s certainly doable. As you’ve seen already, you must pay extremely close attention to each and every detail. *nix can do just about anything, but like any enterprise-class OS, it demands precision. And doing everything bottoms-up as root, one mistake can be catastrophic. But sounds like you’re on your way. Good luck.
Well, sure, as a practical matter those doing the helping need to know what the person being helped already knows. And it’s just good manners to show deference to those who’ve done the work and are willing to help.
But just because a person is new to something doesn’t mean he isn’t able to do it. There are plenty of IT folks with lots of experience who also have lousy judgment, deliver poor quality, never get along with their users, and are resistant to change.
What experience does more than anything is shorten the cycle time, avoids mistakes, and reduces the cost of implementation; obviously all extremely important. And what matters more than how much experience (quantitative) is what was learned from that experience (qualitative).
At the same time, a person with good natural smarts, intellectual curiosity, the ability to read with understanding, humble enough to seek and accept coaching, the ability to work well with colleagues, and the patience and perseverance to see the job through, will also be successful given adequate time.
I spent >2 decades in consulting with a Silicon Valley computer company, and we looked for all of these qualities.