Linux Kernel 3.14 Series Has Reached End of Life (softpedia.com) 99
Slashdot reader prisoninmate quotes an article on Softpedia: it looks like the Linux kernel maintainers decided that there's no need to maintain the Linux kernel 3.14 LTS series anymore, so earlier today, September 11, 2016, they decided to release that last maintenance update, version 3.14.79, and mark the series as EOL (End of Life). Famous Linux kernel maintainer Greg Kroah-Hartman was the one to make the big announcement, and he's urging users who want to still run a long-term supported kernel version to move to the Linux 4.4 LTS series, which is currently the most advanced LTS branch, or use the latest stable release, Linux kernel 4.7.3...
Linux kernel 3.14.79 is a very small update that changes a total of 12 files, with 45 insertions and 17 deletions, thus fixing a bug in the EXT4 file system, a networking issue related to the Reliable Datagram Sockets (RDS) protocol, and updating a few HID, s390, SCSI, networking drivers.
Linux kernel 3.14.79 is a very small update that changes a total of 12 files, with 45 insertions and 17 deletions, thus fixing a bug in the EXT4 file system, a networking issue related to the Reliable Datagram Sockets (RDS) protocol, and updating a few HID, s390, SCSI, networking drivers.
Re: (Score:2)
I've still got systems running 3.13.
WAIT! I've still got systems running 2.6.
If they are doing their jobs - why not?
Re:What? (Score:5, Funny)
Re: What? (Score:3)
What many non support/admin IT types do not realize is we are responsible for security. I've seen write-ups hit those whose IE 6 systems allowed ransomware. Oddly the one doing the write-up denied the request to update.
So damned if you do damned if you don't hence why we hate not forced updates, but rather the security patch is the new agile release etc. I left Mozilla Firefox after 4.0 for this reason and downgraded my office to IE.
With Windows 10 and soon server 2k12 and Windows 7 being only cumulative I
Re: What? (Score:2)
Ok explain the next October update for server 2012 R2 is all or nothing which breaks apps as it includes +110 other updates just to fix one security whole?
No you can't pick and choose hence cumulative
Re: (Score:1)
You handle it just like you would any other update. You put it through your testing pipeline, see what breaks, determine whether you can deploy workarounds or compensating measures for any outstanding defects, and make a final business decision on whether you can proceed to production. If the end result is "nope, we can't deploy this," it's probably time to reevaluate the suitability of that operating system with regard to your business use case. -PCP
Re: (Score:2)
If the end result is "nope, we can't deploy this," it's probably time to reevaluate the suitability of that operating system with regard to your business use case. -PCP
This is just precious. Next explain how you, the erstwhile IT guy, goes about telling the people who decided long ago that they were going to go with the gold standard of the industry, Microsoft - that they have to change their Operating system.
I've related my Vista OS support nightmare many times in here. It involves experts, a shitload of wasted money, dismissals, horrible infighting, and eventual desperation. And an ultimatum by yours truly to fix the system, or wallow in the shithole they put themse
Re: (Score:2)
This is just precious. Next explain how you, the erstwhile IT guy, goes about telling the people who decided long ago that they were going to go with the gold standard of the industry, Microsoft - that they have to change their Operating system.
Easy, in a way they can understand. This would include risk assessments, cost : benefit ratios, short / medium term benefit : inevitable data breach cost ratios, and if necessary lots of pretty charts and graphs if words are too hard for your upper echelon to understand. And documentation that you actually both made the suggestions, and gave detailed data on what could and probably will happen in the case of them not taking suggestions in case they still decide not to follow their IT teams advice, just to c
Re: (Score:2)
This is just precious. Next explain how you, the erstwhile IT guy, goes about telling the people who decided long ago that they were going to go with the gold standard of the industry, Microsoft - that they have to change their Operating system.
Easy, in a way they can understand.
And when after your logical impressive and factual presentation, you are told "We are an institution that uses Windows, because it is the industry standard, and we will continue to use the industry standard. You can go now".
One of my favorite industry standard moments was years ago, when we had a Video Toaster demo for the entire organization. In the day, the toaster of course, ran on the Amiga platform. After wowing the crowd, one of the head honchos said "We have to have this, but it has to be on the Mi
Re: (Score:2)
How in hell do I avoid shell shock, ransomware, but still no releases bringing in bugs all at the same time??
Proper change management and deployment processes, with integrated automated testing (including security auditing/validation)
8 hours of sleep a night, belief in the power of prayer, and probiotics, and a good account with your psychic friend helps too.
Re: (Score:1)
Not really, you're just misunderstanding the whole linux kernel ecosystem.
Unless you're running a plain vanilla 3.14 kernel, this just plain does not affect you.
Most users will be using a distro kernel, and it's the distribution that has the responsibility to maintain their kernel versions, which they will still do.
For example, RHEL 5 still gets security updates until March next year, and runs kernel 2.6.18 (originally released in 2006).
I'm sorry, but your entire post both misses the point, and comes off as
Re: (Score:2)
Umm... security?
If it's a system disconnected from the world with no incoming data possible whatsoever (though I would like to know what the system is good for, then), ok. In all other cases, running an ancient system is probably flirting with disaster.
Re: (Score:2)
Some time ago, I took a shit and now we have a software store in Win10. Coincidence? Correlation? Diarrhea?
Re: (Score:2)
Comment removed (Score:5, Insightful)
Re: (Score:1)
You don't understand kernel development.
For a bug you have, it only takes ONE.
for a second thing, Linux kernels are far more downward compatible than you realize. And WAY more downward compatible than Windows.
Re: (Score:3)
Riiight, and if you need a part for your Chevy here is some iron ore, you CAN smelt iron...right?
Well, an iron ingot is easier to obtain than ore, and obtaining a programming book and learning to hack the kernel is easier and cheaper than obtaining an iron ingot, a machine shop, and learning how to use machining/milling tools.
But that's besides the point. The point is that for IT shops, you can always pay somebody to write a patch for your dated Linux systems, which your existing IT staff can then mass deploy. However if your windows systems go out of support, it's time to buy new computers, servers, e
Re: (Score:2)
"Well, an iron ingot is easier to obtain than ore"
Not true. You can grab dry land just about anywhere and you'll likely have iron oxide (an ore of iron) right there. How much dry land out there will allow youto just reach down and grab an iron ingot?
Re: More reasons to use Windows instead of Linux (Score:2)
Oh that's easy, just walk into a trailer park and you'll easily find some iron scrap sitting around. Sure, it's not an ingot, but who cares? It'll do what you want.
Re: More reasons to use Windows instead of Linux (Score:3)
Hairy I friended you because I agree with 80% of your rants. I used to agree with Linux too with no ABI meaning breaks with updates.
However, Windows 10 is proving us wrong. Next month all versions of Windows both server and desktop will be one cumulative patch?! Also, I am very close to downgrading back to Windows 8.1. Windows 10 breaks ALOT. More than Linux now.
This past month I have seen PowerShell DSC, nested Hyper-V virtualization, update at a later time, and AMD Crimson drivers all break with updates.
Re: (Score:3)
Re: (Score:2)
Long term Hairy that is not a good solution as 78.1 also will be updated the same way. Eventually Windows 7 will go EOL in a few 3 short years.
AMD/Intel will only work with Windows 10 in the newer chips coming out and the world moves on. Of course you can help upgrade which makes you more money at your shop :-)
I think Linux LTS might be the only way to go or a Mac if the user really needs Adobe, auto cad, or Office which is what is tying them to win32. I hope MS does improve QA as they desperately need it.
Re: (Score:2)
Re: (Score:1)
Riiight, and if you need a part for your Chevy here is some iron ore, you CAN smelt iron...right?
That analogy is for hardware not software. Anyway, why would I want to buy a Chevy in the first place?
And people wonder why MSFT can put out 3 stinkers in a row and get more users in a fricking week than Linux has gotten for their desktop in 22 years
Since you are talking about the Desktop you are quite correct but do you really stop and ask why?
I supposed you have never heard of what most people in the IT industry call the "Microsoft Tax" and the fact that most people think that a PC, tablet or "smart phone" are a commodity and could not care less about the operating system and the underlying kernel. Actually my bad I probably should not have used "s
Re: (Score:2)
Anyway, why would I want to buy a Chevy in the first place?
Because it's not a Ford, of course!
Re: (Score:2)
Riiight, and if you need a part for your Chevy here is some iron ore, you CAN smelt iron...right?
Well, sure you can if you like, but you'd probably be better off hiring a mechanic like the rest of us. Or, you know, pick a distro with 10 years support if you want that sort of thing.
Re: (Score:2, Interesting)
At least with Windows, you know how long your OS is going to be maintained for. They don't just randomly decide that they're done issuing updates and discontinue it. Besides, each version of Windows is maintained longer than Linux, which is another plus. And unlike Linux, Windows still supports old hardware like 32 bit systems. There's no reason to replace a system that's working just fine, but that's what your forced to do when developers just randomly decide that it's time to stop maintaining a version of Linux.
As a longtime Linux user, it would make me very happy if you (and everyone else who thinks like you do) would stay on Windows.
Beating Microsoft at the desktop game is a puerile pissing contest that I don't care about. However I do care about interacting with a community of users who can understand the most basic and easily Googled things, such as the ease with which the Linux kernel can be used on a 32-bit system. See how that works?
I agree with Alvinrod that you are probably a troll, but who knows? Ther
Re: (Score:1)
This is quote from another AC:
All Linux LTS releases have an EOL schedule, which is public, and known at least one year in advance. For 3.14, the projected EOL was August/2016. GregKH even gave it an extra month...
Current list:
https://www.kernel.org/category/releases.html
Longterm release kernels Version, Maintainer, Released, Projected EOL
4.4 Greg Kroah-Hartman 2016-01-10 Feb, 2018
4.1 Sasha Levin 2015-06-21 Sep, 2017
3.18 Sasha Levin 2014-12-07 Jan, 2017
3.16 Ben Hutchings 2014-08-03 Apr, 2020
3.12 Jiri Slaby 2013-11-03 Jan, 2017
3.10 Willy Tarreau 2013-06-30 Oct, 2017
3.4 Li Zefan 2012-05-20 Sep, 2016
3.2 Ben Hutchings 2012-01-04 May, 2018
Really, this is linked right there on the front page of www.kernel.org ("Releases" link on top of page).
So YES, we know how long Linux kernels are going to be maintained for.
YES, You can also find most of big distributions' support periods from their websites.
and YES, Linux kernel still supports 32 bit systems and more.
Support period may vary between distributions though, most of them support upgrade to newer versions for free! So theoretically Linux supports a lot longer than each version of Windows. You are obviously a troll. I wonder if you have ever used Linux or BSD before.
Re: (Score:3)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
Re: (Score:1)
No they will not fully support Windows 7 until the advertised EOL. At least not on Kaby Lake processors, which are due out in a few months. Windows 7 will not be officially supported on Kaby Lake at all. Windows 7 is only supported to the advertised EOL if you continue to run it on pre-Kaby-Lake CPUs.
Thus many corporates have had to adjust their plans to keep running Windows 7.
I run the latest Skylake chipset and my Z170 motherboard only supports Windows 8.and 10 as well as "other OS" which works fine with my Linux (Fedora 23 and now 24) distribution. Even Mint works flawlessly in a virtual machine. I also have Windows 10 genuine malware edition which you can get here [microsoft.com] also in a virtual machine, however I don't like the way it likes to talk on the internet (Wireshark is your friend here) even though I have locked down the system, hence I don't run it.
Re: (Score:2)
That's what the LTS branches are for...and unlike Windows, you're free to upgrade whenever it suits you. Linux runs on 32 bit systems just fine, you just need a better distro.
Re: (Score:2)
The bullshit is strong with this one.
Linux is a kernel, not an OS. With windows, you don't know how long your kernel will be maintained for. They don't tell you, and it gets updated every so often. Who cares?
If you want to know how long your OS is updated for, then, um, look at how long the vendor says it's updated for. For example Ubuntu LTS is updated for 5 years. Redhat Enterprise gets 10 years of support.
And unlike Linux, Windows still supports old hardware like 32 bit systems.
Except, you know, Linux d
Re: (Score:2)
Trolling, ignorant or spreading FUD? Don't know, and frankly, I don't know if I care.
Re: (Score:2)
The kernel maintainers always announce the expected EOL date when they announce a new LTS kernel. This wasn't randomly decided. It was known for years in advance by people who, unlike you, know what they're talking about.
3.14 (Score:5, Funny)
Why, if it is now EOL, and they are releasing the last update, did they not make it "159"
Then we could run Pi Linux 3.14.159
I know, I know, version numbers and such, this was still a missed opportunity.
Re: (Score:2)
No, just keep adding new digits after 3.14 FOREVER! ;)
Re: (Score:2)
No, just keep adding new digits after 3.14 FOREVER! ;)
Isn't that too \TeX{}nical for Linux geeks??
--
E
Re: (Score:2)
Technical for nerds? Oh brother.
It was EOLed right on schedule (Score:4, Informative)
All Linux LTS releases have an EOL schedule, which is public, and known at least one year in advance. For 3.14, the projected EOL was August/2016. GregKH even gave it an extra month...
Current list:
https://www.kernel.org/category/releases.html
Longterm release kernels Version, Maintainer, Released, Projected EOL
4.4 Greg Kroah-Hartman 2016-01-10 Feb, 2018
4.1 Sasha Levin 2015-06-21 Sep, 2017
3.18 Sasha Levin 2014-12-07 Jan, 2017
3.16 Ben Hutchings 2014-08-03 Apr, 2020
3.12 Jiri Slaby 2013-11-03 Jan, 2017
3.10 Willy Tarreau 2013-06-30 Oct, 2017
3.4 Li Zefan 2012-05-20 Sep, 2016
3.2 Ben Hutchings 2012-01-04 May, 2018
Really, this is linked right there on the front page of www.kernel.org ("Releases" link on top of page).
Re: (Score:1)
Meanwhile, Windows 10 is supported until October 14, 2025.
Re: (Score:1)
Re: (Score:2)
Re: It was EOLed right on schedule (Score:2)
Nope Windows 10 is supported for 6 months. Then you need bi annual update. 9 months for the pro version if you choose to defer updates on the CBB business update channel.
Afterwards you are not supported and shit will break if you do update. I ll take Linux Ubuntu LTS thank you very much ... If I didn't have a need for Windows software sigh
Re: (Score:2)
Nope Windows 10 is supported for 6 months. Then you need bi annual update. 9 months for the pro version if you choose to defer updates on the CBB business update channel.
Afterwards you are not supported and shit will break if you do update. I ll take Linux Ubuntu LTS thank you very much ... If I didn't have a need for Windows software sigh
wrong.
CB is 4 months minimum
CBB is 8 months minimum
LTSB is 10 years.
Re: It was EOLed right on schedule (Score:2)
Right but it still is 4 to 8 months of support as cumulative updates break apps!
So unless I want to rent my computer for the enterprise edition is there really 10 years of support ... As in just security fixes or bug fixes and not feature updates or mini platform changes? Is the LTSB really 10 years or will new things break shit all the time?
Re: (Score:2)
Re: (Score:2)
Sight difference. The "support", at least the LTS, for Windows usually only includes patches that keeps it running and secure. Support for new hardware or hardware features usually ends MUCH earlier. This is also why Windows 7, which still gets security updates, will never support any of the new features in the next CPU generation. And the same applies to new features for graphics cards, and of course that also means that if there will be a next-gen of USB you will only have a chance for it being supported
LTR isn't all that long (Score:5, Insightful)
Looking at the still-supported LTR kernels, even the oldest one isn't all that old.
For network-connected embedded systems (routers, network-connected printers, IoT, etc.), I would want a kernel that had security-but-patch-maintenance for at least the useful life of the hardware itself - 5-10 years in most cases, longer in some cases like cars, refrigerators, etc.
Perhaps we need a "Very long term release" with an expectation of least 10 years of security fixes, at least for the architecture-independent systems and for a short list of "VLTS-supported" architectures and devices that are in common use in embedded systems.
As someone pointed out already, at least with Linux the source code is available. On the other hand, Microsoft does give 10 years of support for Windows 7/8/8.1 and for the "Long term service branch" mode of Windows 10.
Re: (Score:1)
2.4 kernel was mantained up to 2012, perfect for your WRT54GL.
Re:LTR isn't all that long (Score:4, Interesting)
If you pay it, they will come.
I hope I don't give away a trade secret here, but there are actually companies that took it upon themselves to add years to the LTS. RHEL 5 LTS EOL will be in March 2017, RHEL 6 in November 2020. Those are Kernel 2.6 systems, and we're still talking about patches that add support for new hardware on top of security fixes. Support for security patches only (i.e. no added support for new hardware, pretty much where you're already with pretty much all non-current Windows versions) will end in November 2020 for RHEL 5. We don't know yet when RHEL 6 will reach this final EOL, but expect it for June 2024.
Re: (Score:1)
Re: (Score:1)
There is. 2.6.32 had around 6 years of support. [kernel.org] Looks like 3.16 will be supported until 2020. [kernel.org] Longer would be nice, sure, but 5-6 years for long term seems reasonable for a not-paying-for-it product.
When you install a particular Linux distribution it will come with a Linux kernel version number. Over time you should be updating your particular Linux distribution and this means updating the kernel as well so as that distribution reaches "end of life" it's kernel will not be the same kernel the original distribution started with. This is no more different to what Microsoft does with it's products.
Why do you want longer? Linux distribution are normally free unless you wish to play for a service contract
Re: (Score:1)
The latest RHELv6 runs on 2.6.32 -- just because Kernel.Org won't supporting it, doesn't mean that commercial vendors do not. RHELv6 is supported to late 2020.
Re: (Score:1)
The original question I replied to stated they were thinking about IoT devices and embedded systems like routers. There tends to not be many officially supported distros for those kinds of markets, rendering everyone's discussion on distro support policies moot.
Re: (Score:1)
Choosing a distro is hardly a Sisyphean task. Most people settle on Ubuntu or Mint after doing 5 minutes research and they'll do just fine. Seriously, having so many options is the single biggest strength of Linux. The only people who obsess about fragmentation are bean counters employed at Redmond or Cupertino. Those kinds of people see the world as big market instead of a collection of interesting groups and opportunities.
Don't panic (Score:5, Informative)
Re: (Score:2)
Yes, it's a bit weird that someone posted up "we need an LTS with 10 years of fixes" when there's been a very prominent one (Red Hat and its derivatives) available for quite some time. RHEL isn't targeted at embedded systems admittedly and they've only just introduced a preview of an ARM variant with version 7. The oldest supported RHEL kernel is the creaking 2.6.18 - with a ton of backports - found in RHEL 5, but beware that support for that ends in 6 months...
Re: (Score:2)
Well, there's support and there's support. Backporting will probably end in March 17, but you will get critical security fixes for RHEL5 until 2020.
It is advisable, though, that you migrate away from it before that. Preferably even before March 17.
Re: (Score:2)
That's not exactly true ... that 's only if you specifically pay for the Extended Lifecycle Support subscription (which incidentally still supports RHEL4) ...
Red Hat Support Policy [redhat.com]
Without that specific extra subscription RHEL5 goes EOL March 2017... and CentOS5 will get the last of its updates then as they don't get and rebuild the ELS packages.
Re: (Score:2)
Looking at the still-supported LTR kernels, even the oldest one isn't all that old.
For network-connected embedded systems (routers, network-connected printers, IoT, etc.), I would want a kernel that had security-but-patch-maintenance for at least the useful life of the hardware itself - 5-10 years in most cases, longer in some cases like cars, refrigerators, etc.
If you are designing a device like this and care about maintaining for security then you need to have a continual, preferably over the air, upgrade system in place.
Given that you would be looking at several libc updates over that period upgrading a kernel major version should be expected and not a significant problem. You would be much better off investing more in the update and recovery system than backporting kernel security fixes for ten years.
Re: (Score:2)
For network-connected embedded systems (routers, network-connected printers, IoT, etc.), I would want a kernel that had security-but-patch-maintenance for at least the useful life of the hardware itself - 5-10 years in most cases, longer in some cases like cars, refrigerators, etc.
If the device is under warranty then you would expect support, replacement or reimbursement in the event of failure. Once something is out of support you can continue to use it and even organize third party maintenance but the company is under no obligation to continue to provide support.
Perhaps we need a "Very long term release" with an expectation of least 10 years of security fixes, at least for the architecture-independent systems and for a short list of "VLTS-supported" architectures and devices that are in common use in embedded systems
The software and all the tools are available and well documented. Knock yourself out. Who knows you may make some money.
As someone pointed out already, at least with Linux the source code is available. On the other hand, Microsoft does give 10 years of support for Windows 7/8/8.1 and for the "Long term service branch" mode of Windows 10.
You can get long-term support with Linux if you pay for a service contract and that is the same for Mic
Re: (Score:2)
For network-connected embedded systems (routers, network-connected printers, IoT, etc.), I would want a kernel that had security-but-patch-maintenance for at least the useful life of the hardware itself - 5-10 years in most cases, longer in some cases like cars, refrigerators, etc.
Why? When was the last time you actually upgraded a kernel by hand? That's a very rare thing to do. If you use a distro, then they'll provide stable kernels for the life of the distro, for 5 to 10 years and the upgrade will happen
Re: (Score:2)
Liar! Nobody ever finishes compiling a kernel, the moment it is done, the moment it is outdated!
renumbering? (Score:2)
It's a pity, for this last release, then didn't break with the numbering convention and just jump to 3.14.159...
Re: (Score:2)
It's a pity, for this last release, then didn't break with the numbering convention and just jump to 3.14.159...
Why do something as pointless as that?
Most current Linux distributions are running a version 4 kernel. My Fedora 24 distribution is running "kernel-core-4.7.2-201.fc24.x86_64" as per the time of me writing, however the initial kernel release was kernel-core-4.5.5-300. I am still running Fedora 24 but the underlying kernel and appropriate packages that needed updating have been updated. So unless you are directly involved with maintaining software who cares if the version numbers have been changed.
Actual
Do you hear that Google? (Score:2)
I'm sick of all my Android devices running woefully outdated kernels. Get on it!