Fedora "Beefy Miracle" Comes To IBM Mainframes

[phoronix]

Phoronix: Fedora "Beefy Miracle" Comes To IBM Mainframes

Fedora 17, the latest release that is cooked with an odd codename, has just been released for IBM System z...

http://www.phoronix.com/vr.php?view=MTEyOTc

[uid313]

Nobody is ever going to run a bleeding-edge distribution like Fedora on a mainframe anyway.

Mainframes run tried-and-proven enterprise LTS distributions.

[jvillain]

Ya, I was wonder where there could possibly be a market for this with the possible exception of some old abandoned piece of hardware in the corner of a server room. But any piece of hardware like that would be better being scrapped any way. Odd that they would do this while at the same time fighting over whether to support arm.

[RahulSundaram]

Ya, I was wonder where there could possibly be a market for this with the possible exception of some old abandoned piece of hardware in the corner of a server room. But any piece of hardware like that would be better being scrapped any way. Odd that they would do this while at the same time fighting over whether to support arm.

Regardless of what impression you might get out of reading Phoronix and the opinions expressed in a random mailing list thread, there is no real fight about Fedora ARM. Fedora ARM has existed for years and will likely become primary architecture in the next release or two. It is just a matter of time and nothing more.

As for Fedora for mainframes, it does make sense since Fedora is upstream for Red Hat Enterprise Linux and continuation of development in a public manner is in line with free and open source development and Red Hat can incorporate feedback. There are developers out there running Fedora on mainframes to make sure the latest open source stuff works well on it and things like Anaconda changes coming up in Fedora 18 don't cause any regressions on that architecture.

[FireBurn]

Ya, I was wonder where there could possibly be a market for this with the possible exception of some old abandoned piece of hardware in the corner of a server room. But any piece of hardware like that would be better being scrapped any way. Odd that they would do this while at the same time fighting over whether to support arm.

Have you even checked how much money IBM makes from Mainframe? Mainframe is vital - it rarely goes wrong - when it does go wrong everyone knows about it

[allquixotic]

Have you even checked how much money IBM makes from Mainframe? Mainframe is vital - it rarely goes wrong - when it does go wrong everyone knows about it

I work closely with mainframes as part of my day job. Trust me, they go wrong. Most of what goes wrong is due to human/programming error at the application level, though; it's rarely the hardware or the OS's fault. Still, they aren't particularly stable because the complicated applications that run on them are chock full of bugs.

[popper]

I work closely with mainframes as part of my day job. Trust me, they go wrong. Most of what goes wrong is due to human/programming error at the application level, though; it's rarely the hardware or the OS's fault. Still, they aren't particularly stable because the complicated applications that run on them are chock full of bugs.

LOL OC they are, it not like the worlds scientists are known for actually bench marking never mind optimizing and re-factorizing their code for speed, conformity and that published paper result is all they care about, after all if they need more speed they just put in another requisition for more nodes to be added to the core mainframe, that's the whole point of mainframes and the masses of cash AMD,NV, Intel, IBM etc chase year on year, how else are they going to finance that next CPU/GPU Co-processor tick/tock

[allquixotic]

LOL OC they are, it not like the worlds scientists are known for actually bench marking never mind optimizing and re-factorizing their code for speed, conformity and that published paper result is all they care about, after all if they need more speed they just put in another requisition for more nodes to be added to the core mainframe, that's the whole point of mainframes and the masses of cash AMD,NV, Intel etc chase year on year, how else are they going to finance that next CPU/GPU Co-processor tick/tock

Actually my work is in operations, not research... very large-scale WWW environment... but still, issues

[DeepDayze]

Nobody is ever going to run a bleeding-edge distribution like Fedora on a mainframe anyway.

Mainframes run tried-and-proven enterprise LTS distributions.

Mainframes can serve as hosts for virtualized OS's too you know and they can handle many thousands of VM's on one host as well as making use of the high performance IO subsystems that mainframes have.

[DeepDayze]

I work closely with mainframes as part of my day job. Trust me, they go wrong. Most of what goes wrong is due to human/programming error at the application level, though; it's rarely the hardware or the OS's fault. Still, they aren't particularly stable because the complicated applications that run on them are chock full of bugs.

Even zOS can have its issues as well

I've worked in a mainframe shop as an operator and have dealt with abends that are due to bugs within the OS.