The linked page claims that CloudLinux has pledged $1M/yr to AlmaLinux which solves one of the main concerns I have about Rocky.
It also seems like CloudLinux has already been doing something like this internally so it shouldn't be an issue to get the project up and going. It will be interesting to see how they do vs Rocky.
There's a difference between a community backed effort vs a CEO of a small company that wants to throw $1M at something. Considering a community would do it for free, I'm under the impression that the CEO just wants to make money.
I view the risk differently: A community is self-supporting and contributors are equal.
If one entity throws 1M/yr onto it, they could suddenly stop and the contributors go away and no other contributors can take over as they never were equals ...
(And yes, it's more complicated, first case can also lose key contributors who have keys etc.)
> Launched with the code-name Project Lenix, AlmaLinux is an open-source, community-driven project that intends to fill the gap left by the demise of the CentOS stable release. AlmaLinux is a 1:1 binary compatible fork of RHEL® 8 and it is built by the creators of the established CloudLinux OS.
This is what open source is all about. Personally I don't need it, but I love it.
That's not the same thing, it's a whole different thing to do a "rpm --rebuild *.src.rpm" and to actually take a upstream source package it with a spec, patch it to make it compatible with other things in the distro and test them in different setup.
The challenge anyway is not to rebuild packages or make packages it is to actually maintain the build farms and the build tools.
source: i work for Red Hat (not doing any packages work tho) and I used to work for Mandrakesoft on Linux-Mandrake when we were basically rebuild Red Hat rpms to our flavour.
I totally get that, but I meant more in the philosophical sense. For instance many other build farms in the world have compiled the linux kernel before it got to redhat.
Yeah, this is always the problem with apocalypse-level events like this. There's always a period of fragmentation where everyone and their Aunt Lillian starts a new fork project to pursue whatever angle they want.
It'll take some time for all the fragments to coalesce, likely with some T1000-ing into larger projects while others lose steam as developers settle on a single option. Unfortunately, in the meantime consumers of CentOS are left to look at the field in the crystal ball and try to figure out which project to place their bet on (or just move to something else for a while and have a pint while the whole thing blows over). The good news is that things don't really become pressing until 2023, which gives some time for things to resolve.
It's very hard to make strategic decisions on CentOS replacements that don't actually exist. Looking at the Rocky forums, there's far more energy complaining about the name than seemingly in assisting any release.
I was just starting towards a migration from CentOS 7 to CentOS 8 when Red Hat made their announcement so I put that on hold and decided to stick with 7 for now.
In about six months or so, I'll look at which of these new RHEL 8 clones actually have something to show off and start making a decision from there. No point in chasing something which is currently vaporware and may still be half a year from now.
This is great news, we have been using CloudLinux and Kernel Care at my web hosting company for many years and they are great products. These guys know what they are doing and I trust this distro will be well supported.
18 comments
[ 3.1 ms ] story [ 55.1 ms ] threadIt also seems like CloudLinux has already been doing something like this internally so it shouldn't be an issue to get the project up and going. It will be interesting to see how they do vs Rocky.
If one entity throws 1M/yr onto it, they could suddenly stop and the contributors go away and no other contributors can take over as they never were equals ...
(And yes, it's more complicated, first case can also lose key contributors who have keys etc.)
This is what open source is all about. Personally I don't need it, but I love it.
The challenge anyway is not to rebuild packages or make packages it is to actually maintain the build farms and the build tools.
source: i work for Red Hat (not doing any packages work tho) and I used to work for Mandrakesoft on Linux-Mandrake when we were basically rebuild Red Hat rpms to our flavour.
It'll take some time for all the fragments to coalesce, likely with some T1000-ing into larger projects while others lose steam as developers settle on a single option. Unfortunately, in the meantime consumers of CentOS are left to look at the field in the crystal ball and try to figure out which project to place their bet on (or just move to something else for a while and have a pint while the whole thing blows over). The good news is that things don't really become pressing until 2023, which gives some time for things to resolve.
It's the Circle Of Open Source. :)
This is so overblown. People now get "minor" version patches before RHEL instead of after and now people go nuts.
This is crazy. I'm sure only a handful of people really needs to follow after RHEL and not slightly before it.
In that sense, Fedora is considered civilization killer instability compared to CentOS Stream.
I was just starting towards a migration from CentOS 7 to CentOS 8 when Red Hat made their announcement so I put that on hold and decided to stick with 7 for now.
In about six months or so, I'll look at which of these new RHEL 8 clones actually have something to show off and start making a decision from there. No point in chasing something which is currently vaporware and may still be half a year from now.