openSUSE Tumbleweed team changes its mind about x86-64-v2
(2022/12/16)
- Reference: 1671210070
- News link: https://www.theregister.co.uk/2022/12/16/tumbleweed_reverses_x864v2_plan/
- Source link:
Tumbleweed is changing course once again, but it's due to popular demand, and it means broader compatibility for more people. Saying that, it's looking for someone to help maintain its 32-bit support.
Back in November, the openSUSE project [1]announced that it was changing the minimum CPU requirement for Tumbleweed on x86-64 to version 2 of the instruction set – as well as [2]dropping support for x86-32 machines .
This would bring Tumbleweed into line with the requirements of its next-gen enterprise OS, [3]currently available as a prototype called ALP . There's a [4]summary of the issue on the project's website. However, following an outcry from openSUSE users, the plan has changed… again. It's still dropping the x86-32 edition, but it's not going to switch to requiring second-generation x86-64 chips.
[5]
Tumbleweed release manager [6]Dominique Leuenberger , who publishes the regular Tumbleweed "Review of the Week" [7]blog , told The Reg :
The backlash was way larger than expected – and we do love our users and want to keep them.
So instead of moving the entire distro to v2 (or anything but baseline), we started investing much more into hwcaps and figuring out which libraries benefit the most of being built for v3, making them parallel installed (and auto-picked by glibc ). A bit more wasted disk space, but it will work for old and new machines.
The hwcaps [8]feature in glibc allows detection and manipulations of the h ard w are cap abilitie s of chips in various CPU families, which Tumbleweed has supported [9]for a while .
Tumbleweed is taking on increased importance. Its main corporate sponsor SUSE is [10]aiming its next-gen enterprise distro towards an immutable root filesystem and containerized workloads. That pulls the rug out from underneath openSUSE Leap, which is the current stable-release version of openSUSE. Leap releases [11]have been synchronized with SLE since 15.3 , which means that if SLE is replaced by ALP, Leap no longer has a base to draw from.
[12]openSUSE makes baseline CPU requirements a little friendlier than feared
[13]Strong support for Snap and Ubuntu Core as Canonical meet IRL
[14]Heavy, man: Tuxedo puts out 2.2kg Stellaris AMD Gen 4
[15]Canonical makes Ubuntu Pro free for up to five machines
So, Tumbleweed is the future of openSUSE. It does have some unique advantages over other rolling-release distros such as Arch Linux. Before Tumbleweed's software packages become part of the distro, they undergo automated testing via SUSE's [16]openQA tool. Only when they pass are they incorporated. Then, if some unanticipated problem with some bleeding-edge component manifests, openSUSE's integrated support for Btrfs snapshots means that changes can be rolled back, returning the OS to a previous state. It's somewhat akin to Windows' [17]System Restore feature, and it's missing from the Ubuntu and Red Hat families.
Some issues around dropping the x86-32 edition still remain, though. Although the 32-bit edition is being demoted to a "port", someone still needs to maintain the 32-bit compatibility packages, and as the last set of [18]meeting minutes show, someone needs to step up to do this. SUSE Prague engineer [19]posted a "Call for ix86 maintainers" on the Factory mailing list.
[20]
We hope someone volunteers. The package for the [21]Munin monitoring software also [22]needs a new caretaker. ®
Get our [23]Tech Resources
[1] https://news.opensuse.org/2022/11/28/tw-to-roll-out-mitigation-plan-advance-microarchitecture/
[2] https://www.theregister.com/2022/11/30/opensuse_drops_x86_32/
[3] https://www.theregister.com/2022/10/05/suse_alp_v001/
[4] https://en.opensuse.org/openSUSE:X86-64-Architecture-Levels
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/oses&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Y5z4jV0fJ0MzDrB35cHMnQAAANY&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[6] https://people.opensuse.org/2016-11-25-dominique-leuenberger.html
[7] https://dominique.leuenberger.net/blog/
[8] https://www.gnu.org/software/libc/manual/html_node/Hardware-Capability-Tunables.html
[9] https://hackweek.opensuse.org/20/projects/support-glibc-hwcaps-and-micro-architecture-package-generation
[10] https://www.theregister.com/2022/07/26/suse_adaptable_linux_platform/
[11] https://www.theregister.com/2021/06/04/opensuse_leaps_to_153_now/
[12] https://www.theregister.com/2022/11/30/opensuse_drops_x86_32/
[13] https://www.theregister.com/2022/11/09/canonical_conference/
[14] https://www.theregister.com/2022/11/08/tuxedo_stellaris_amd_gen_4/
[15] https://www.theregister.com/2022/10/06/ubuntu_pro_free/
[16] http://open.qa/
[17] https://support.microsoft.com/en-us/windows/use-system-restore-a5ae3ed9-07c4-fd56-45ee-096777ecd14e
[18] https://lists.opensuse.org/archives/list/factory@lists.opensuse.org/thread/LALV24HMN6HWGQDTXSUOKL2O747H4YVB/
[19] https://lists.opensuse.org/archives/list/factory@lists.opensuse.org/thread/LALV24HMN6HWGQDTXSUOKL2O747H4YVB/
[20] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/oses&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Y5z4jV0fJ0MzDrB35cHMnQAAANY&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[21] https://munin-monitoring.org/
[22] https://lists.opensuse.org/archives/list/factory@lists.opensuse.org/thread/NFUGPRNRVW6DJB4UUL33NMGEY433RTXR/
[23] https://whitepapers.theregister.com/
Back in November, the openSUSE project [1]announced that it was changing the minimum CPU requirement for Tumbleweed on x86-64 to version 2 of the instruction set – as well as [2]dropping support for x86-32 machines .
This would bring Tumbleweed into line with the requirements of its next-gen enterprise OS, [3]currently available as a prototype called ALP . There's a [4]summary of the issue on the project's website. However, following an outcry from openSUSE users, the plan has changed… again. It's still dropping the x86-32 edition, but it's not going to switch to requiring second-generation x86-64 chips.
[5]
Tumbleweed release manager [6]Dominique Leuenberger , who publishes the regular Tumbleweed "Review of the Week" [7]blog , told The Reg :
The backlash was way larger than expected – and we do love our users and want to keep them.
So instead of moving the entire distro to v2 (or anything but baseline), we started investing much more into hwcaps and figuring out which libraries benefit the most of being built for v3, making them parallel installed (and auto-picked by glibc ). A bit more wasted disk space, but it will work for old and new machines.
The hwcaps [8]feature in glibc allows detection and manipulations of the h ard w are cap abilitie s of chips in various CPU families, which Tumbleweed has supported [9]for a while .
Tumbleweed is taking on increased importance. Its main corporate sponsor SUSE is [10]aiming its next-gen enterprise distro towards an immutable root filesystem and containerized workloads. That pulls the rug out from underneath openSUSE Leap, which is the current stable-release version of openSUSE. Leap releases [11]have been synchronized with SLE since 15.3 , which means that if SLE is replaced by ALP, Leap no longer has a base to draw from.
[12]openSUSE makes baseline CPU requirements a little friendlier than feared
[13]Strong support for Snap and Ubuntu Core as Canonical meet IRL
[14]Heavy, man: Tuxedo puts out 2.2kg Stellaris AMD Gen 4
[15]Canonical makes Ubuntu Pro free for up to five machines
So, Tumbleweed is the future of openSUSE. It does have some unique advantages over other rolling-release distros such as Arch Linux. Before Tumbleweed's software packages become part of the distro, they undergo automated testing via SUSE's [16]openQA tool. Only when they pass are they incorporated. Then, if some unanticipated problem with some bleeding-edge component manifests, openSUSE's integrated support for Btrfs snapshots means that changes can be rolled back, returning the OS to a previous state. It's somewhat akin to Windows' [17]System Restore feature, and it's missing from the Ubuntu and Red Hat families.
Some issues around dropping the x86-32 edition still remain, though. Although the 32-bit edition is being demoted to a "port", someone still needs to maintain the 32-bit compatibility packages, and as the last set of [18]meeting minutes show, someone needs to step up to do this. SUSE Prague engineer [19]posted a "Call for ix86 maintainers" on the Factory mailing list.
[20]
We hope someone volunteers. The package for the [21]Munin monitoring software also [22]needs a new caretaker. ®
Get our [23]Tech Resources
[1] https://news.opensuse.org/2022/11/28/tw-to-roll-out-mitigation-plan-advance-microarchitecture/
[2] https://www.theregister.com/2022/11/30/opensuse_drops_x86_32/
[3] https://www.theregister.com/2022/10/05/suse_alp_v001/
[4] https://en.opensuse.org/openSUSE:X86-64-Architecture-Levels
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/oses&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Y5z4jV0fJ0MzDrB35cHMnQAAANY&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[6] https://people.opensuse.org/2016-11-25-dominique-leuenberger.html
[7] https://dominique.leuenberger.net/blog/
[8] https://www.gnu.org/software/libc/manual/html_node/Hardware-Capability-Tunables.html
[9] https://hackweek.opensuse.org/20/projects/support-glibc-hwcaps-and-micro-architecture-package-generation
[10] https://www.theregister.com/2022/07/26/suse_adaptable_linux_platform/
[11] https://www.theregister.com/2021/06/04/opensuse_leaps_to_153_now/
[12] https://www.theregister.com/2022/11/30/opensuse_drops_x86_32/
[13] https://www.theregister.com/2022/11/09/canonical_conference/
[14] https://www.theregister.com/2022/11/08/tuxedo_stellaris_amd_gen_4/
[15] https://www.theregister.com/2022/10/06/ubuntu_pro_free/
[16] http://open.qa/
[17] https://support.microsoft.com/en-us/windows/use-system-restore-a5ae3ed9-07c4-fd56-45ee-096777ecd14e
[18] https://lists.opensuse.org/archives/list/factory@lists.opensuse.org/thread/LALV24HMN6HWGQDTXSUOKL2O747H4YVB/
[19] https://lists.opensuse.org/archives/list/factory@lists.opensuse.org/thread/LALV24HMN6HWGQDTXSUOKL2O747H4YVB/
[20] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/oses&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Y5z4jV0fJ0MzDrB35cHMnQAAANY&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[21] https://munin-monitoring.org/
[22] https://lists.opensuse.org/archives/list/factory@lists.opensuse.org/thread/NFUGPRNRVW6DJB4UUL33NMGEY433RTXR/
[23] https://whitepapers.theregister.com/
Tumbleweed is the future of openSUSE.
"Tumbleweed is taking on increased importance. Its main corporate sponsor SUSE is aiming its next-gen enterprise distro towards an immutable root filesystem and containerized workloads. That pulls the rug out from underneath openSUSE Leap, which is the current stable-release version of openSUSE. Leap releases have been synchronized with SLE since 15.3, which means that if SLE is replaced by ALP, Leap no longer has a base to draw from."
Seems the author doesn't really understand the relation of the various SUSE Linux distributions.
The source is Factory, where the packages are built and QA'd. Once they pass, they go into Tumbleweed. Certain snapshots of Tumbleweed which are mature enough then form the basis for both, SUSE Enterprise Linux (SEL) and openSUSE Leap.
Essentially, Tumbleweed is to SUSE what Fedora is to Red Hat, and openSUSE is what CentOS has been before it was "repurposed" by Red Hat.
ALP, which is one potential(!) candidate for the next major version of both SUSE Linux Enterprise and openSUSE Leap. ALP is also based on MicroOS, which sits between Tumbleweed and SEL/Leap (although closer to the latter). Should ALP end up becoming the next major version of SEL and its openSUSE Leap equivalent (whatever it will be named), which is still highly questionable as no decision has been made (at this time, ALP is more like an experiment with uncertain future), then these versions will still sit behind Tumbleweed (which might also be named differently by then).
In any case, SEL 15 and with it openSUSE Leap still have many more years to come, and even if the decision is made to progress with ALP then it will take several years before becomes a real product. So any concerns about the future of openSUSE Leap at this point are pointless, and whatever the next version of SEL and openSUSE LTS will be there for sure will be an easy migration path once SEL 15 and Leap 15 become EOL.