I came late to linux in heavy use, being a BSD person of old. Commercialized BSD exists but the foundation(s) behind the major flavours are clearly NFP models.
So as a late linux adopter, in Debian, I am wondering why people see value in SUSE and RedHat since all my experience has been the SLA value is .. dreck. No problem we had with a RH kernel or OS behaviour was fixed inside our $SLA$ in the EL release we had. All problems were either WONTFIX or re-install later EL.
Debian update/upgrade appears to work. Fixes get applied. What am I missing? If they want to own SAP-ON-LINUX or ORACLE-ON-LINUX why should I worry?
I am more concerned about GPL and ZFS these days, than commercial backing: I dont think it actually adds value to me. (20+ nodes in a research cluster, cloud)
A lot of it has to do with standardization and training. If you go with Red Hat, for example, you get a bunch of certifications and training materials and standardized tools like Satellite and so on and so forth. It's easier to find solutions and peace of mind when you get some guidance from a vendor. It's similar to managing a network with Windows technologies. Many of the choices are already made for you.
If you go with Debian you have to spend more time building your own stack which is fine if you're experienced and you know what you're doing or if you're the lone IT guy. But if you build a stack for a team then you have to make sure that your employees and new hires know not only the specific stack that you built but also how to find support for each of the individual components.
If you go with a heterogeneous environment that consists of several distributions like a mix of Red Hat, Debian, Ubuntu, Arch, etc., then you'll be tempted to jump out of a window.
These sort of decisions are made at a level where they want "support".
A C(I|T)O wants someone that they feel that they can shout at if something goes wrong.
For you, with 20 nodes managed by yourself or a smaller team, CentOS / openSUSE / Debian / Fedora / Ubuntu all work.
(They also work for much larger situations, but RedHat / SUSE will sell execs on "fleet management tools" and "integrated installers" "SSO" and access to the devs that wrote the software running on top, and extract a fee. Most companies would be better employing a few highly skilled people, but that comes out of a different budget)
One of the things companies like SUSE and RH provide is testing across a variety of expensive storage and compute platforms. It turns out that continuous integration testing is just getting started in kernel land. Major Hayden had a cool presentation at TXLF 19 about it.
I can't comment on the value of SLAs, as I was never directly involved (but what I heard is pretty much what you mention). What I can tell you, however, is that there are companies where higher-ups won't agree to an IT proposal if it doesn't include commercial support, whether useful or not. Company policy mandates it so it gets done.
Yea. Probably thats how we wound up there. We bought off risk, by spending money that it turned out didn't resolve the risk when it hit, but because we TRIED we got to shift blame, by not owning "you broke it you fix it" to the management view.
I know people that work at SuSE, in a nutshell, they are like Redhat in that they don't really provide linux only, they provide ways of integrating it into your business.
That is where Debian just isn't going to compete. And that is why SuSE exists, they're selling much more than linux.
> But right now, I think we're all staring at alpine in kubernetes asking if the old giant unix model is the one we want.
I think "all" is a bit of an exaggeration though :). Besides, something is hosting all those Alpine containers, and while I'm pretty far removed from this field, I hear it's not Alpine most of the time.
- Redhat and Suse have LONG binary compatibility, like 20 years. Debian has 3 or 5 years?
- If something goes wrong, you can sue and get something. Debian has no money.
- Redhat and Suse actually develops linux kernel, underlying filesystems etc... Very important if you have weird concurrent database issue on large cluster.
- Debian is political and has no commercial predictability. For example there is very small chance it may decide to abandon systemd and move to something else.
- Commercial certification; you wont get SAP R/3 or Oracle support contract for Debian, even if would manage to install it.
> - Redhat and Suse actually develops linux kernel, underlying filesystems etc... Very important if you have weird concurrent database issue on large cluster.
That is what they sell you alright, but actually getting access to these devs, unless you are a massive customer (and even then it can be difficult) you will never see anything more than a bug report, which might actually get more traction in an upstream bug tracker
Couldn't give you an answer to that. I think it is curious that many developers go the debian route, while a lot of network administrators use centOS.
Maybe network administrators value the option to switch to a system where SLAs are more common?
I think SUSE is just used by people who always used it. I doubt they have as much pull to newcomers as other distros. Personally I never did like yast. I don't know if that is still in use to be honest.
So as a late linux adopter, in Debian, I am wondering why people see value in SUSE and RedHat since all my experience has been the SLA value is .. dreck. No problem we had with a RH kernel or OS behaviour was fixed inside our $SLA$ in the EL release we had. All problems were either WONTFIX or re-install later EL.
Debian update/upgrade appears to work. Fixes get applied. What am I missing? If they want to own SAP-ON-LINUX or ORACLE-ON-LINUX why should I worry?
I am more concerned about GPL and ZFS these days, than commercial backing: I dont think it actually adds value to me. (20+ nodes in a research cluster, cloud)