I wasn’t aware it was higher than one, this is pretty cool. However, despite the size of the bone they’re tossing, I don’t think this is enough to shake the image of a company trying to cash out its good will. I’m very skeptical of using EL for the foreseeable future, including the rebuilds.
The only aspect I would be iffy on (mainly because I've never looked) are the things you get just for being a Red Hat subscription holder. These are the Red Hat hosted services, like the Customer Portal and Hybrid Cloud Console (Insights, etc).
But the actual products you get from the subscriptions? Yeah, those are all open source software with available upstream.
As for value, the hosted stuff is actually pretty cool and helpful for getting proactive analysis, suggestions, alerts, etc of your environments. In general using the Red Hat product over the upstream provides a handful of benefits. Usually the software is pretty well stabilized, you're not getting random updates that require massive changes. Deployment can also be easier than upstream, or at least better documented. For example, deploying and maintaining AAP is a breeze, and can be done on traditional servers instead of k8s. It provides all the disparate pieces in one place, rather than needing you to deploy AWX, Galaxy-ng, Pinakes, Keycloak, etc on your own manually.
The other non-software pieces, like the knowledge base, discussions, docs, hosted apps, etc are all effectively bonus material of a sort. But they are really good bonus material to have.
RedHat creates and drives development of many OSS projects. FreeIPA, strimzi, systemd are some examples. Anybody can use these projects without using the RedHat labeled version. But none of this projects would be well maintained without RedHat.
> It is a program and an offering designed for individual developers, available through the Red Hat Developer program.
On the previous conversation about this change, it was was pointed out that developers have to be individuals, and that if you are an organisation that uses RedHat you can't use this.
However I haven't been able to get anyone to confirm what the developer cost would be in that situation or if it is still free.
I think that will probably be more relevant for a lot of organisations that have some RHEL on their prod boxes but don't want to spend so much getting it on their dev/test environments.
License is valid only for a year after which you need to re-register. IBM might one day decide to put an end to it and you would have to either pay up or migrate all your servers. After they killed off CentOS and Rocky/Alma i wouldn't trust them.
OK, this is time. We need new RedHat - distro/firm that DO development and pays for it. So earns too...
Which non-rpm based distro could be base for this ? Becouse rpm-based distros are (except RH) are pure joke, a re-compilators. Debian if you want to start from scratch, Ubuntu could be if they finally stard developing source code instead of looking something to sell. Not Arch, not Void - not business mentality. Maybe Void as a base after revriting everything into musl, including things like Oracle ;)
NixOs is in best technical possition but so far just sources (big) set and no comparable to RH development expertise...
Actually no one have RH-like expertise as far as I know. So new business-level distro needs funding for 15 years or so ?
"Back to the basics!" then :) Use stable, backcompatible, futureproof technologies ! C, C++, Perl (hehe he), git and maybe finally Lua, D and Julia ? In US Ruby probably too. Drop non-sense things like Gnome and moust of WWW insanity.
Or just wait what new IBM will bring to rest of us... Like Winland with it's by-design over-complications and imposibilities to have stable env.
- they are re-compilators. With long history and didn't sold in recent years, yes, but still small compared to RH. You can hear they do things, have customers not only from Germany so they have a lot of expertise but to my nose they are not realy self-sufficient without RH. They are more like small mechanic shop close to global factory - low throughput. Am I hand waving too much ? It's becouse SUSE pro&cons are not realy known. So moust likely they are small shop. Even if 2nd after RH.
- they use rpm's so with a bit of pressure or becouse of management wanting to cover asses they or their customers will easilly roll back to IBM
- SUSE do not look like wanting to be global. Not saying it is easy or even possible without controlling your own cpu's production...
Also:
- I asked for non-rpm distros
- I personally am burned out with rpms since 90's - blame RH for not implementing Debian like net features years after Debian did it
- git repo is better binary package. Source tarballs for LTS versions, done. IMO. And tar can be replaced by whatever in "tarbals"...
> Red Hat and SUSE have been around doing a lot, and it's a shame that so many people don't want to pay for actual commercial Linux expertise. :(
Agree. But easy to understood: a) MS is monopoly, Apple do not count as platform for development (multi-billion corporations don't count and even them do not develop on Apple - just modern "terminals" :) ) b) no serious RH alternative - as in: security and open source code development, including new features and apps
> they use rpm's so with a bit of pressure or becouse of management wanting to cover asses they or their customers will easilly roll back to IBM
That's a weird take. Moving to a different package format is usually trivial. If you're using fpm, it may be as simple as a single option. It's the change to anything in the environment that causes issues. Different services / configs / defaults etc. is what gets you when packaging.
> Moving to a different package format is usually trivial
It is not. It is more like tear down and rebuild everything. By humans. And env/configs on the top. Look on it from management perspective - sometimes zero technical knowledge but a lot marketing assurance of effortless continuation.
Because "format" is not the only variable, it realy means whole corporation install base. Even if you have build farms ready.
Just imagine that discussions: "There is only one way to be sure..." ;)
Sure it is a rebuild of everything, but the build steps don't change almost at all, just the boilerplate around them. Seriously, I've done a few distros, changing the format is the trivial part. You can even covert many end results automatically https://help.ubuntu.com/community/RPM/AlienHowto The environment is the main/only problem.
Anyone that doesnt offer support is meaningless. Redhat's costumer base is business who need linux but need support, and 27/7 with a 4h-SLA. Anyone who doesn't offer this is dead in the water as far as poaching RHEL's costumer base. In fact, companies care very little about open-source, they just want linux with support. Previously paying $600 per server per year would otherwise be stupid when CentOS existed.
21 comments
[ 4.8 ms ] story [ 60.9 ms ] threadAgreed. RHEL is the replacement cash cow for AIX.
They can't milk it for as much profit if others can get milk elsewhere.
https://developers.redhat.com/articles/faqs-no-cost-red-hat-...
* RHEL
* RHEL EUS (stick on a minor version for 2 years)
* Ansible Automation Platform
* A slew of Red Hat middleware solutions
Or is the value in being able to access the documentation for things, which is behind a login-wall these days?
But the actual products you get from the subscriptions? Yeah, those are all open source software with available upstream.
As for value, the hosted stuff is actually pretty cool and helpful for getting proactive analysis, suggestions, alerts, etc of your environments. In general using the Red Hat product over the upstream provides a handful of benefits. Usually the software is pretty well stabilized, you're not getting random updates that require massive changes. Deployment can also be easier than upstream, or at least better documented. For example, deploying and maintaining AAP is a breeze, and can be done on traditional servers instead of k8s. It provides all the disparate pieces in one place, rather than needing you to deploy AWX, Galaxy-ng, Pinakes, Keycloak, etc on your own manually.
The other non-software pieces, like the knowledge base, discussions, docs, hosted apps, etc are all effectively bonus material of a sort. But they are really good bonus material to have.
On the previous conversation about this change, it was was pointed out that developers have to be individuals, and that if you are an organisation that uses RedHat you can't use this.
However I haven't been able to get anyone to confirm what the developer cost would be in that situation or if it is still free.
I think that will probably be more relevant for a lot of organisations that have some RHEL on their prod boxes but don't want to spend so much getting it on their dev/test environments.
Which non-rpm based distro could be base for this ? Becouse rpm-based distros are (except RH) are pure joke, a re-compilators. Debian if you want to start from scratch, Ubuntu could be if they finally stard developing source code instead of looking something to sell. Not Arch, not Void - not business mentality. Maybe Void as a base after revriting everything into musl, including things like Oracle ;)
NixOs is in best technical possition but so far just sources (big) set and no comparable to RH development expertise...
Actually no one have RH-like expertise as far as I know. So new business-level distro needs funding for 15 years or so ?
"Back to the basics!" then :) Use stable, backcompatible, futureproof technologies ! C, C++, Perl (hehe he), git and maybe finally Lua, D and Julia ? In US Ruby probably too. Drop non-sense things like Gnome and moust of WWW insanity.
Or just wait what new IBM will bring to rest of us... Like Winland with it's by-design over-complications and imposibilities to have stable env.
[...]
> Actually no one have RH-like expertise as far as I know.
So what is SUSE, chopped liver? They've been around for slightly longer and do a lot upstream too.
Red Hat and SUSE have been around doing a lot, and it's a shame that so many people don't want to pay for actual commercial Linux expertise. :(
I didn't mentioned SUSE because:
- they are re-compilators. With long history and didn't sold in recent years, yes, but still small compared to RH. You can hear they do things, have customers not only from Germany so they have a lot of expertise but to my nose they are not realy self-sufficient without RH. They are more like small mechanic shop close to global factory - low throughput. Am I hand waving too much ? It's becouse SUSE pro&cons are not realy known. So moust likely they are small shop. Even if 2nd after RH.
- they use rpm's so with a bit of pressure or becouse of management wanting to cover asses they or their customers will easilly roll back to IBM
- SUSE do not look like wanting to be global. Not saying it is easy or even possible without controlling your own cpu's production...
Also:
- I asked for non-rpm distros
- I personally am burned out with rpms since 90's - blame RH for not implementing Debian like net features years after Debian did it
- git repo is better binary package. Source tarballs for LTS versions, done. IMO. And tar can be replaced by whatever in "tarbals"...
> Red Hat and SUSE have been around doing a lot, and it's a shame that so many people don't want to pay for actual commercial Linux expertise. :(
Agree. But easy to understood: a) MS is monopoly, Apple do not count as platform for development (multi-billion corporations don't count and even them do not develop on Apple - just modern "terminals" :) ) b) no serious RH alternative - as in: security and open source code development, including new features and apps
That's a weird take. Moving to a different package format is usually trivial. If you're using fpm, it may be as simple as a single option. It's the change to anything in the environment that causes issues. Different services / configs / defaults etc. is what gets you when packaging.
It is not. It is more like tear down and rebuild everything. By humans. And env/configs on the top. Look on it from management perspective - sometimes zero technical knowledge but a lot marketing assurance of effortless continuation.
Because "format" is not the only variable, it realy means whole corporation install base. Even if you have build farms ready.
Just imagine that discussions: "There is only one way to be sure..." ;)