6:55 a.m.
Subject: Are Meta/Facebook servers using Fedora Linux?
Good day from Singapore,
I have just come across this article.
Article: Fedora's FESCo Rejects The Idea Of "-fno-omit-frame-pointer"
As Default Compiler Flag
Link: https://www.phoronix.com/news/Fedora-Rejects-No-Omit-FP
[QUOTE]
As a change proposal first initiated by Meta/Facebook developers, they
wanted -fno-omit-frame-pointer and -mno-omit-leaf-frame-pointer to be
added to the default C/C++ compilation flags.
...snipped...
Meta engineers believe that any performance cost is small and worth it
while SUSE engineers previously cited around possible 5~10%
regressions.
[/QUOTE]
From the above quotes, I thought that Meta/Facebook servers are using
Fedora Linux, or at least Linux servers.
Anyone can confirm?
Regards,
Mr. Turritopsis Dohrnii Teo En Ming
Targeted Individual in Singapore
Blogs:
https://tdtemcerts.blogspot.com
https://tdtemcerts.wordpress.com
7:27 a.m.
On Tue, Dec 6, 2022 at 7:56 AM Turritopsis Dohrnii Teo En Ming
<tdtemccnp(a)gmail.com> wrote:
Subject: Are Meta/Facebook servers using Fedora Linux?
Good day from Singapore,
I have just come across this article.
Article: Fedora's FESCo Rejects The Idea Of "-fno-omit-frame-pointer"
As Default Compiler Flag
Link: https://www.phoronix.com/news/Fedora-Rejects-No-Omit-FP
[QUOTE]
As a change proposal first initiated by Meta/Facebook developers, they
wanted -fno-omit-frame-pointer and -mno-omit-leaf-frame-pointer to be
added to the default C/C++ compilation flags.
...snipped...
Meta engineers believe that any performance cost is small and worth it
while SUSE engineers previously cited around possible 5~10%
regressions.
[/QUOTE]
From the above quotes, I thought that Meta/Facebook servers are using
Fedora Linux, or at least Linux servers.
Anyone can confirm?
I don't have first hand knowledge of Meta, but I have a fair amount of
enterprise experience. A large enterprise will have a large number of
technology stacks. It will look like the wild, wild west. They will
likely allow Fedora, in addition to other distros like Alpine, Debian,
the BSDs and Ubuntu. And they will also allow a heap of developmental
languages.
I often recommend Fedora Server anytime I see folks using RHEL or
CentOS. I don't understand why organizations run that antique software
that is no longer in development. Fedora provides modern software and
is in active development with continuous bug fixes.
The "in active development" part is important. Old versions of
software and kernels just accumulate more unfixed bugs over time. Most
developers don't spend time on old versions of software, so the known
bugs don't get fixed. Adversaries love that property of old software.
https://thenewstack.io/design-system-can-update-greg-kroah-hartman-linux-...
.
I eat my own dog food. I run Fedora Servers at my house.
Jeff
9:53 a.m.
If you aren't buying support from the OS distributor and/or the application
developer (and the app developer providing the support will only support
the enterprise releases) then it makes little sense to use the "enterprise"
variant.
If you have to stay up on patches then the enterprise OSes might have some
value if you were not committed to keeping fedora on a version still
getting updates.
If you aren't paying for support and/or you are providing all of the
support internally, then you might as well be on one of the Fedora like
leading edge distributions with all the new features.
On Tue, Dec 6, 2022 at 7:27 AM Jeffrey Walton <noloader(a)gmail.com> wrote:
On Tue, Dec 6, 2022 at 7:56 AM Turritopsis Dohrnii Teo En Ming
<tdtemccnp(a)gmail.com> wrote:
>
> Subject: Are Meta/Facebook servers using Fedora Linux?
>
> Good day from Singapore,
>
> I have just come across this article.
>
> Article: Fedora's FESCo Rejects The Idea Of "-fno-omit-frame-pointer"
> As Default Compiler Flag
> Link: https://www.phoronix.com/news/Fedora-Rejects-No-Omit-FP
>
> [QUOTE]
>
> As a change proposal first initiated by Meta/Facebook developers, they
> wanted -fno-omit-frame-pointer and -mno-omit-leaf-frame-pointer to be
> added to the default C/C++ compilation flags.
>
> ...snipped...
>
> Meta engineers believe that any performance cost is small and worth it
> while SUSE engineers previously cited around possible 5~10%
> regressions.
>
> [/QUOTE]
>
> From the above quotes, I thought that Meta/Facebook servers are using
> Fedora Linux, or at least Linux servers.
>
> Anyone can confirm?
I don't have first hand knowledge of Meta, but I have a fair amount of
enterprise experience. A large enterprise will have a large number of
technology stacks. It will look like the wild, wild west. They will
likely allow Fedora, in addition to other distros like Alpine, Debian,
the BSDs and Ubuntu. And they will also allow a heap of developmental
languages.
I often recommend Fedora Server anytime I see folks using RHEL or
CentOS. I don't understand why organizations run that antique software
that is no longer in development. Fedora provides modern software and
is in active development with continuous bug fixes.
The "in active development" part is important. Old versions of
software and kernels just accumulate more unfixed bugs over time. Most
developers don't spend time on old versions of software, so the known
bugs don't get fixed. Adversaries love that property of old software.
https://thenewstack.io/design-system-can-update-greg-kroah-hartman-linux-...
.
I eat my own dog food. I run Fedora Servers at my house.
Jeff
_______________________________________________
users mailing list -- users(a)lists.fedoraproject.org
To unsubscribe send an email to users-leave(a)lists.fedoraproject.org
Fedora Code of Conduct:
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives:
https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org
Do not reply to spam, report it:
https://pagure.io/fedora-infrastructure/new_issue
1:16 a.m.
On Tue, 6 Dec 2022 at 23:54, Roger Heflin <rogerheflin(a)gmail.com> wrote:
If you aren't buying support from the OS distributor and/or the application developer
(and the app developer providing the support will only support the enterprise releases)
then it makes little sense to use the "enterprise" variant.
If you have to stay up on patches then the enterprise OSes might have some value if you
were not committed to keeping fedora on a version still getting updates.
If you aren't paying for support and/or you are providing all of the support
internally, then you might as well be on one of the Fedora like leading edge distributions
with all the new features.
I do not pay for support most of the time. I can resolve most issues
using Google search. If Google search cannot help me, I will ask
questions in the mailing lists and community forums.
Regards,
Mr. Turritopsis Dohrnii Teo En Ming
Targeted Individual in Singapore
>
>
> On Tue, Dec 6, 2022 at 7:27 AM Jeffrey Walton <noloader(a)gmail.com> wrote:
>>
>> On Tue, Dec 6, 2022 at 7:56 AM Turritopsis Dohrnii Teo En Ming
>> <tdtemccnp(a)gmail.com> wrote:
>> >
>> > Subject: Are Meta/Facebook servers using Fedora Linux?
>> >
>> > Good day from Singapore,
>> >
>> > I have just come across this article.
>> >
>> > Article: Fedora's FESCo Rejects The Idea Of
"-fno-omit-frame-pointer"
>> > As Default Compiler Flag
>> > Link: https://www.phoronix.com/news/Fedora-Rejects-No-Omit-FP
>> >
>> > [QUOTE]
>> >
>> > As a change proposal first initiated by Meta/Facebook developers, they
>> > wanted -fno-omit-frame-pointer and -mno-omit-leaf-frame-pointer to be
>> > added to the default C/C++ compilation flags.
>> >
>> > ...snipped...
>> >
>> > Meta engineers believe that any performance cost is small and worth it
>> > while SUSE engineers previously cited around possible 5~10%
>> > regressions.
>> >
>> > [/QUOTE]
>> >
>> > From the above quotes, I thought that Meta/Facebook servers are using
>> > Fedora Linux, or at least Linux servers.
>> >
>> > Anyone can confirm?
>>
>> I don't have first hand knowledge of Meta, but I have a fair amount of
>> enterprise experience. A large enterprise will have a large number of
>> technology stacks. It will look like the wild, wild west. They will
>> likely allow Fedora, in addition to other distros like Alpine, Debian,
>> the BSDs and Ubuntu. And they will also allow a heap of developmental
>> languages.
>>
>> I often recommend Fedora Server anytime I see folks using RHEL or
>> CentOS. I don't understand why organizations run that antique software
>> that is no longer in development. Fedora provides modern software and
>> is in active development with continuous bug fixes.
>>
>> The "in active development" part is important. Old versions of
>> software and kernels just accumulate more unfixed bugs over time. Most
>> developers don't spend time on old versions of software, so the known
>> bugs don't get fixed. Adversaries love that property of old software.
>>
https://thenewstack.io/design-system-can-update-greg-kroah-hartman-linux-...
>> .
>>
>> I eat my own dog food. I run Fedora Servers at my house.
>>
>> Jeff
6:03 p.m.
On Dec 6, 2022, at 08:27, Jeffrey Walton <noloader(a)gmail.com> wrote:
I often recommend Fedora Server anytime I see folks using RHEL or
CentOS. I don't understand why organizations run that antique software
that is no longer in development. Fedora provides modern software and
is in active development with continuous bug fixes.
The "in active development" part is important. Old versions
of
software and kernels just accumulate more unfixed bugs over time. Most
developers don't spend time on old versions of software, so the known
bugs don't get fixed. Adversaries love that property of old software.
That a pretty ignorant view of RHEL/CentOS.
Red Hat backports known bug fixes to the software in RHEL. It also has modules that are
updated at a faster cadence, too.
I realize this is a Fedora list, but this kind of misinformation doesn’t really help
Fedora. Many companies aren’t really interested in significant rearchitecting core parts
of the service because of Fedora’s rapid pace. Just because it is the newest version
doesn’t mean it has a backwards compatible API. New software also has new bugs, and less
testing.
I do agree that Fedora Server is a powerful platform, and I use it myself. But I’d have a
hard time arguing it makes sense for customers running enterprise services with project
lifetimes extending for several years. Making them use Fedora would often result in
companies running EOL versions, leaving a platform with even more security holes.
Both Fedora and RHEL/CentOS have their place.
--
Jonathan Billings
6:30 p.m.
On Tue, Dec 6, 2022 at 7:04 PM Jonathan Billings <billings(a)negate.org> wrote:
On Dec 6, 2022, at 08:27, Jeffrey Walton <noloader(a)gmail.com> wrote:
> I often recommend Fedora Server anytime I see folks using RHEL or
> CentOS. I don't understand why organizations run that antique software
> that is no longer in development. Fedora provides modern software and
> is in active development with continuous bug fixes.
> The "in active development" part is important. Old versions of
> software and kernels just accumulate more unfixed bugs over time. Most
> developers don't spend time on old versions of software, so the known
> bugs don't get fixed. Adversaries love that property of old software.
That a pretty ignorant view of RHEL/CentOS.
I don't believe it's as ignorant as you think. Unfortunately, it's the
reality of the situation.
Red Hat backports known bug fixes to the software in RHEL. It also
has modules that are updated at a faster cadence, too.
The keyword is "known". Developers don't work on 5 or 10 year old
software. Existing bugs don't get uncovered and fixed. They don't
become "known", so they don't get backported and fixed.
That's exactly the point Greg HK makes at
https://thenewstack.io/design-system-can-update-greg-kroah-hartman-linux-...
. No one is working on old kernels. Problems in old kernels are going
to remain latent and unfixed. Kernel developers and Red Hat may
backport an occasional fix, but they are not fixing all the problems.
The same thing happens in userland. I help maintain several projects,
and contribute to many others. None of those projects concern
themselves with 5 or 10 year old software. The problems in old
software remain unfixed.
The way to avoid the problems is to use something that's being
actively developed. That's what Greg HK is telling us. Fedora makes it
especially easy to use software that is being actively developed
because Fedora usually packages the current release of whatever is
going into the package.
And dnf-system-upgrade every 6 months is a small price to pay for
Fedora (https://docs.fedoraproject.org/en-US/quick-docs/dnf-system-upgrade/).
You get Red Hat processes and stability with modern software. Its a
win-win.
Jeff
7:40 p.m.
On Dec 6, 2022, at 19:32, Jeffrey Walton <noloader(a)gmail.com> wrote:
And dnf-system-upgrade every 6 months is a small price to pay for
Fedora (https://docs.fedoraproject.org/en-US/quick-docs/dnf-system-upgrade/).
You get Red Hat processes and stability with modern software. Its a
win-win.
I suppose it’s a win if you’re the one being paid to fix broken systems. :/
I am fine with your suggestion for the modern containerized platforms or cloud service,
using agile methodology, and constantly spinning out new versions. Release early and
release often. Not everyone has that workflow.
It all boils down to risk assessment.
Greg K-H has a good point and what he describes is one of the risks with using enterprise
Linux. The kernel in RHEL doesn’t get all fixes in the upstream. On one hand, one of those
bug fixes might end up being a security hole down the road. But on the other hand, that is
a risk that Red Hat takes and is responsible to fix. This is why you pay for RHEL.
In your model, every bug, breaking api change and orphaned package is your problem and
it’s up to you to fix whatever code you use to run your service.
Some companies would rather pay another company to take that risk than shoulder it
themselves.
I’m glad that Fedora is the basis of future RHEL. I will be able to tell what will work
and will not in new releases. I’ll get valuable experience in the new software and maybe
even get a hand in guiding the path to future releases. I will have the same basic OS on
my laptop as what I’m running in production. I am also happy to know that people use
Fedora Server in production as an alternative to other distros.
I just don’t agree with your assessment that enterprise Linux has no value. Not everyone
has the headcount or technical ability to use Fedora for their infrastructure. And I think
it is disingenuous to suggest otherwise.
--
Jonathan Billings
10:46 p.m.
On Tue, 2022-12-06 at 19:30 -0500, Jeffrey Walton wrote:
The keyword is "known". Developers don't work on 5 or
10 year old
software. Existing bugs don't get uncovered and fixed. They don't
become "known", so they don't get backported and fixed.
That's exactly the point Greg HK makes at
https://thenewstack.io/design-system-can-update-greg-kroah-hartman-linux-...
. No one is working on old kernels. Problems in old kernels are going
to remain latent and unfixed. Kernel developers and Red Hat may
backport an occasional fix, but they are not fixing all the problems.
Odd. On my CentOS 7 system, my last kernel update was mid-November,
the prior the month before, another a month before that, another a
month before that. Hardly not being worked on. The kernel is an on-
going thing, and various distros backport what they can.
And we do have applications with long history, LibreOffice being just
one that springs to mind. Though some people might argue some projects
never fix bugs, as many bug reports go back many years with no
resolutions, it's full speed ahead and damn the torpedoes.
The same thing happens in userland. I help maintain several
projects,
and contribute to many others. None of those projects concern
themselves with 5 or 10 year old software. The problems in old
software remain unfixed.
My experience with programmers is that they don't like having the rug
pulled out from under them mid-development. There are many programs
that spend years in development (pre- and post-release). That's harder
to do when you have to keep re-learning the quirks of systems.
Likewise with sysadmins. Many features upgrades are entirely unwanted,
bug fixes is all they're interested in.
--
uname -rsvp
Linux 3.10.0-1160.80.1.el7.x86_64 #1 SMP Tue Nov 8 15:48:59 UTC 2022 x86_64
Boilerplate: All unexpected mail to my mailbox is automatically deleted.
I will only get to see the messages that are posted to the mailing list.
11:28 p.m.
On Tue, Dec 6, 2022 at 11:46 PM Tim <ignored_mailbox(a)yahoo.com.au> wrote:
On Tue, 2022-12-06 at 19:30 -0500, Jeffrey Walton wrote:
> The keyword is "known". Developers don't work on 5 or 10 year old
> software. Existing bugs don't get uncovered and fixed. They don't
> become "known", so they don't get backported and fixed.
>
> That's exactly the point Greg HK makes at
> https://thenewstack.io/design-system-can-update-greg-kroah-hartman-linux-...
> . No one is working on old kernels. Problems in old kernels are going
> to remain latent and unfixed. Kernel developers and Red Hat may
> backport an occasional fix, but they are not fixing all the problems.
Odd. On my CentOS 7 system, my last kernel update was mid-November,
the prior the month before, another a month before that, another a
month before that. Hardly not being worked on. The kernel is an on-
going thing, and various distros backport what they can.
You are not getting all the fixes. From Greg's talk:
Kroah-Hartman gave a good example of how a benign looking
bug can turn into something nasty. Some three years ago a
TTY1 layer bug was fixed. Back then it looked like a
normal, harmless bug. The kernel community fixed it, pushed
it in the new release and was done with it. Three years
later someone was found it to be a security bug. Now
companies freaked out. They didn’t bother to use the
patches back then, and now even companies like SUSE and Red
Hat had to go back and fix all their old stuff.
Here's one I have first hand knowledge of from 2021. It should have
gotten a CVE, but Mitre's website would not accept my report/request:
https://github.com/weidai11/cryptopp/issues/1072 . Here's another one
from 2018: https://github.com/weidai11/cryptopp/issues/602 .
I'm fairly certain an attacker can exploit both of those. Neither got
CVEs (though I tried). Neither was backported to earlier versions of
the library.
I've seen other projects do the same with memory errors. Some don't
even bother trying to get a CVE because they consider it a security
theatre. These bugs are more instances of the Greg's TTY1 layer bug.
No one thinks it is important. Companies like Red Hat and SUSE will
never know to pick-up the change because there's no signalling to
them.
...
> The same thing happens in userland. I help maintain several projects,
> and contribute to many others. None of those projects concern
> themselves with 5 or 10 year old software. The problems in old
> software remain unfixed.
My experience with programmers is that they don't like having the rug
pulled out from under them mid-development. There are many programs
that spend years in development (pre- and post-release). That's harder
to do when you have to keep re-learning the quirks of systems.
No one likes revisiting old code. It's just something you have to do,
like going to the dentist.
Likewise with sysadmins. Many features upgrades are entirely
unwanted,
bug fixes is all they're interested in.
Ideally, you stay lock-step with technology you depend on, and the
break/fix cycle is small as you keep pace with things. But put it off
for several years, and you'll find you boxed yourself into a bad
place. You will find you are carrying a lot of technological debt, and
you are accepting a lot of risk.
And the worst thing about it (to me) is: a company puts off the
maintenance to keep profits up and shareholders happy. But they screw
me, you, our families and friends when the data breach comes. And to
add insult to injury, then they issue those repulsive statements like,
"it was an advanced persistent threat" or "we care deeply about
security." Hogwash. They made their choices, and it was to trade
security and our online safety for profits.
Jeff
1:25 a.m.
On Wed, 7 Dec 2022 at 08:04, Jonathan Billings <billings(a)negate.org> wrote:
On Dec 6, 2022, at 08:27, Jeffrey Walton <noloader(a)gmail.com> wrote:
> I often recommend Fedora Server anytime I see folks using RHEL or
> CentOS. I don't understand why organizations run that antique software
> that is no longer in development. Fedora provides modern software and
> is in active development with continuous bug fixes.
> The "in active development" part is important. Old versions of
> software and kernels just accumulate more unfixed bugs over time. Most
> developers don't spend time on old versions of software, so the known
> bugs don't get fixed. Adversaries love that property of old software.
That a pretty ignorant view of RHEL/CentOS.
Red Hat backports known bug fixes to the software in RHEL. It also has modules that are
updated at a faster cadence, too.
Thank you for this piece of information.
Regards,
Mr. Turritopsis Dohrnii Teo En Ming
Targeted Individual in Singapore
>
> I realize this is a Fedora list, but this kind of misinformation doesn’t really help
Fedora. Many companies aren’t really interested in significant rearchitecting core parts
of the service because of Fedora’s rapid pace. Just because it is the newest version
doesn’t mean it has a backwards compatible API. New software also has new bugs, and less
testing.
>
> I do agree that Fedora Server is a powerful platform, and I use it myself. But I’d
have a hard time arguing it makes sense for customers running enterprise services with
project lifetimes extending for several years. Making them use Fedora would often result
in companies running EOL versions, leaving a platform with even more security holes.
>
> Both Fedora and RHEL/CentOS have their place.
>
> --
> Jonathan Billings
1:36 a.m.
The answer is Hell No!
Fedora is no where near stable enough and secure enough.
On 1/26/2023 11:25 PM, Turritopsis Dohrnii Teo En Ming wrote:
> On Wed, 7 Dec 2022 at 08:04, Jonathan Billings <billings(a)negate.org> wrote:
>> On Dec 6, 2022, at 08:27, Jeffrey Walton <noloader(a)gmail.com> wrote:
>>> I often recommend Fedora Server anytime I see folks using RHEL or
>>> CentOS. I don't understand why organizations run that antique software
>>> that is no longer in development. Fedora provides modern software and
>>> is in active development with continuous bug fixes.
>>> The "in active development" part is important. Old versions of
>>> software and kernels just accumulate more unfixed bugs over time. Most
>>> developers don't spend time on old versions of software, so the known
>>> bugs don't get fixed. Adversaries love that property of old software.
>> That a pretty ignorant view of RHEL/CentOS.
>>
>> Red Hat backports known bug fixes to the software in RHEL. It also has modules
that are updated at a faster cadence, too.
> Thank you for this piece of information.
>
> Regards,
>
> Mr. Turritopsis Dohrnii Teo En Ming
> Targeted Individual in Singapore
>
>> I realize this is a Fedora list, but this kind of misinformation doesn’t really
help Fedora. Many companies aren’t really interested in significant rearchitecting core
parts of the service because of Fedora’s rapid pace. Just because it is the newest version
doesn’t mean it has a backwards compatible API. New software also has new bugs, and less
testing.
>>
>> I do agree that Fedora Server is a powerful platform, and I use it myself. But
I’d have a hard time arguing it makes sense for customers running enterprise services with
project lifetimes extending for several years. Making them use Fedora would often result
in companies running EOL versions, leaving a platform with even more security holes.
>>
>> Both Fedora and RHEL/CentOS have their place.
>>
>> --
>> Jonathan Billings
> _______________________________________________
> users mailing list -- users(a)lists.fedoraproject.org
> To unsubscribe send an email to users-leave(a)lists.fedoraproject.org
> Fedora Code of Conduct:
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives:
https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org
> Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
10:04 a.m.
I often recommend Fedora Server anytime I see folks using RHEL or
CentOS. I don't understand why organizations run that antique software
that is no longer in development. Fedora provides modern software and
is in active development with continuous bug fixes.
The answer is very easy. If it is not broken, do not touch it. There is
nothing SysAdmins hate more than having to solve problems brought by
unnecessary updates to software. Wonder why many banks and businesses
still use COBOL ?
--
------------------------------
/\_/\
|O O| pepebuho(a)gmail.com
~~~~ Javier Perez
~~~~ While the night runs
~~~~ toward the day...
m m Pepebuho watches
from his high perch.
1:11 a.m.
On Tue, 6 Dec 2022 at 21:27, Jeffrey Walton <noloader(a)gmail.com> wrote:
On Tue, Dec 6, 2022 at 7:56 AM Turritopsis Dohrnii Teo En Ming
<tdtemccnp(a)gmail.com> wrote:
>
> Subject: Are Meta/Facebook servers using Fedora Linux?
>
> Good day from Singapore,
>
> I have just come across this article.
>
> Article: Fedora's FESCo Rejects The Idea Of "-fno-omit-frame-pointer"
> As Default Compiler Flag
> Link: https://www.phoronix.com/news/Fedora-Rejects-No-Omit-FP
>
> [QUOTE]
>
> As a change proposal first initiated by Meta/Facebook developers, they
> wanted -fno-omit-frame-pointer and -mno-omit-leaf-frame-pointer to be
> added to the default C/C++ compilation flags.
>
> ...snipped...
>
> Meta engineers believe that any performance cost is small and worth it
> while SUSE engineers previously cited around possible 5~10%
> regressions.
>
> [/QUOTE]
>
> From the above quotes, I thought that Meta/Facebook servers are using
> Fedora Linux, or at least Linux servers.
>
> Anyone can confirm?
I don't have first hand knowledge of Meta, but I have a fair amount of
enterprise experience. A large enterprise will have a large number of
technology stacks. It will look like the wild, wild west. They will
likely allow Fedora, in addition to other distros like Alpine, Debian,
the BSDs and Ubuntu. And they will also allow a heap of developmental
languages.
I often recommend Fedora Server anytime I see folks using RHEL or
CentOS. I don't understand why organizations run that antique software
that is no longer in development. Fedora provides modern software and
is in active development with continuous bug fixes.
The "in active development" part is important. Old versions of
software and kernels just accumulate more unfixed bugs over time. Most
developers don't spend time on old versions of software, so the known
bugs don't get fixed. Adversaries love that property of old software.
https://thenewstack.io/design-system-can-update-greg-kroah-hartman-linux-...
.
I eat my own dog food. I run Fedora Servers at my house.
Jeff
We are pampered with too many choices of Enterprise Linux distros.
Regards,
Mr. Turritopsis Dohrnii Teo En Ming
Targeted Individual in Singapore
10:33 a.m.
On 2022-12-06 04:55, Turritopsis Dohrnii Teo En Ming wrote:
From the above quotes, I thought that Meta/Facebook servers are
using
Fedora Linux, or at least Linux servers.
As far as I know, the answer is "No". Their production platform is
based on CentOS Stream.
https://www.youtube.com/watch?v=cA_Nd3crBuA
The presentation as a whole is worth watching for a view of practices in
a large production environment. The migration to Stream is mentioned
around 23:30.
1:20 a.m.
On Wed, 7 Dec 2022 at 00:33, Gordon Messmer <gordon.messmer(a)gmail.com> wrote:
On 2022-12-06 04:55, Turritopsis Dohrnii Teo En Ming wrote:
> From the above quotes, I thought that Meta/Facebook servers are using
> Fedora Linux, or at least Linux servers.
As far as I know, the answer is "No". Their production platform is
based on CentOS Stream.
https://www.youtube.com/watch?v=cA_Nd3crBuA
The presentation as a whole is worth watching for a view of practices in
a large production environment. The migration to Stream is mentioned
around 23:30.
Thanks for the Youtube link. Now I know Meta/Facebook is using CentOS Stream.
Regards,
Mr. Turritopsis Dohrnii Teo En Ming
Targeted Individual in Singapore
2:19 p.m.
On 6 Dec 2022, at 12:55, Turritopsis Dohrnii Teo En Ming
<tdtemccnp(a)gmail.com> wrote:
Meta engineers believe that any performance cost is small and worth it
while SUSE engineers previously cited around possible 5~10%
regressions.
meta engineers do a lot of performance work and their experience is what is being referred
to.
Indeed along with the SUSE engineers experience.
As someone else said they are not using Fedora.
Barry
453
days inactive
505
days old
15 comments
9 participants
participants (9)
-
Barry Scott -
Gordon Messmer -
Javier Perez -
Jeffrey Walton -
Jonathan Billings -
Roger Heflin -
Thomas Dineen -
Tim -
Turritopsis Dohrnii Teo En Ming