Slashdot Mirror
CentOS Linux 6.8 Released (softpedia.com)
An anonymous reader writes: CentOS team is pleased to announce the immediate availability of CentOS Linux 6.8 and install media for i386 and x86_64 Architectures. Release Notes for 6.8 are available here. Softpedia writes: "CentOS Linux 6.8 arrives today with major changes, among which we can mention the latest Linux 2.6.32 kernel release from upstream with support for storing up to 300TB of data on XFS filesystems. The VPN endpoint solution implemented in the NetworkManager network connection manager utility is now provided on the libreswan library instead of the Openswan IPsec implementation used in previous release of the OS, and it looks like the SSLv2 protocol has been disabled by default for the SSSD (System Security Services Daemon), which also comes with support for smart cards now." In addition, the new release comes with updated applications, including the LibreOffice 4.3.7 office suite and Squid 3.4 caching and forwarding web proxy, many of which are supporting the Transport Layer Security (TLS) 1.2 protocol, including Git, YUM, Postfix, OpenLDAP, stunnel, and vsftpd. The dmidecode open-source tool now supports SMBIOS 3.0.0, you can now pull kickstart files from HTTPS (Secure HTTP) sources, the NTDp (Network Time Protocol daemon) package has an alternative solution as chrony, SSLv3 has been disabled by default, and there's improved support for Hyper-V.
Considering CentOS 7.x has been out for well over a year...this is just an extending support for the old CentOS 6 line.
They're just following RHEL which announced 6.8 on 5/10...
Karma: Bad
2.6.32 differs so much from modern kernels that trying to cherry-pick fixes leads to anything but stability. I wouldn't touch such a kernel with a 0.015 furlong pole.
The creatures outside looked from Alt-Right to Antifa; but already it was impossible to say which was which.
CentOS/RedHat motto has always been "stability and security". Nowadays however , I don't think there is much difference with Debian, for instance, in terms of stability/security. Furthermore, for those using the desktop release, RH and CentOS are really behind (eg compared to Ubuntu) in terms of ergonomics, utilities and other applications.
Slashdot, fix the reply notifications... You won't get away with it...
2.6.32 differs so much from modern kernels that trying to cherry-pick fixes leads to anything but stability. I wouldn't touch such a kernel with a 0.015 furlong pole.
rhel kernels are the most heavily tested kernels available, really you would trust a new kernel with your company's data?
Hosts have 3 weeks to roll it out:
(Yes, that's right, NSS, not OpenSSL.)
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/6.8_Release_Notes/new_features_security.html
https://www.paypal-knowledge.com/infocenter/index?page=content&id=FAQ1914
Oh wait, I see Paypal backed off on it, they have moved it ahead to June 30th 2017. Previously it was next month, June 17, 2016. Enough people must have complained. This change was going to break tens of thousands of eCommerce websites, if not more.
So hosts have more time now; a whole year. At least CentOS has addressed it now. I expect hosts are going to keep servers provisioned on 6.x on 6.x for as long as possible.
"CentOS 6 is really behind" means that it does not have a bunch of significant recent changes. Which is the definition of stable.
Has CentOS 6 kept up with recent changes? If so, it's not "really behind". If not, it's stable. Pick one.
I would say they've done as advertised, they kept it pretty stable. That happens to be what I want right now. If I wanted cutting-edge, I might use Fedora.
This isn't even extended support. CentOS 5 is in Extended support until March 31st 2017, 6 is mainstream until that date and extended until November 30th 2020, 7 is mainstream until that date in 2020 and extended support until 2024. See: https://wiki.centos.org/About/... for more info and the latest version of different software on the different releases.
The major.minor kernel number for Red Hat 6.0 was chosen based on what was stable when it was released in 2010. Kernel updates since then are reflected in the revision number. Updates after initial release don't change the API, the ABI, or the major.minor parts of the version number. They change the revision number.
> there is a huge difference between modern Ubuntu like releases and CentOS. Even Ubuntu 14.04 (2 years old) is way ahead
I don't necesarily disagree. Let's assume that's right, that Ubuntu has had a lot of updates (changes) and CentOS hasn't. That's what you said, right?
Of course all that new stuff has new APIs and especially new ABIs. The APIs and ABIs of RHEL 6 haven't changed for six years, so it doesn't have all the new shinies. What do you call it when something doesn't change a lot over time, when it pretty much remains the same? For APIs and ABIs, we call that "stable".
Notice the word is neither "good" nor "bad", it's "stable", aka unchanging, remaining the same, reliable.
> but at least show a bit of fairness.
I can't think of anything more fair than stating a plain, objective fact. RHEL doesn't change the interfaces. They are stable. Love it or hate it, it's a fact. What would be UNFAIR would be to lie and say RHEL doesn't provide a stable environment. That's simply untrue as a factual matter, for the sense of the word "stable" that matters for software maintainence.
There are plenty of things not certified to run on CentOS7/RHEL7.
Also, at the risk of massive flamage - systemd.
Two very good reasons to keep on upgrading CentOS6.
Losing my mod points for this. CentOS 6.8 is designed to be binary compatible with RHEL 6.8, which uses the same kernel version. CentOS 6.8 is NOT the latest CentOS version, with CentOS 7.0 taking that honor. You can hardly be 'binary compatible' with a specific version of RHEL unless you use the same kernel version. A lot of mission critical systems rely on CentOS, so please don't flame them. If you want a bleeding edge kernel, use another distro.
I had to install it this morning by typing ''yum update'', it told me that it was going to install 855 MB and prompted me ''Is this ok [y/N]'' -- notice a default answer of no.
This is yet more evidence that RedHat/CentOS is behind the time and not following recent industry practice of bamboozling their users into installing the latest version of the OS whether they want to or not. Should I downgrade to Microsoft Windows so that I can become as exasperated as some of my friends ?
CentOS 6.8 is NOT the latest CentOS version, with CentOS 7.0 taking that honor.
Actually 7.2 is the latest!
Need to type accents and special characters in Windows? Use FrKeys
Or Gnome3. What a load of old rubbish, I must be an idiot to have almost 3 machines running it.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
I strongly suspect RedHat has a shitload of paying customers who have told them in no uncertain terms that if RH drops RHEL6 and goes systemd-only, they'll be moving to a non-systemd distro.
You expect wrong then. A shrill and tiny minority of trolls and whiners do not represent their customer base. RedHat is perfectly fine about continuing with systemd as are their customers.
RHEL is for companies who are risk averse. They don't need the latest features, or want an OS which is constantly changing in unpredictable ways. They want something which is reliable, stable, supported and does what they bought it for in the first place. If a 2.6 kernel is good for that purpose then what's the problem?
RHEL 7.2 is the latest long-term supported production platform from Red Hat. Fedora is the bleeding edge work, and a sign of what will be in future RHEL releases. The bleeding edge versions of perl, python, and of virtualization toolkits and security toolkits can be very destabilizing to production systems, which is why RHEL and CentOS have been so popular for production work.
I still have a Debian Squeeze system that can run perfectly happily using Debian's 2.6.32 kernel. As it happens, I did experience one (and only one) problem with this system & kernel, which is that USB3 didn't work properly (onboard Etron controller), so I installed Debian's 3.2 backport Sarge kernel and all has been well ever since. I don't suppose lack of USB3 is really going to worry the average production server operator (the target market for RedHat/CentOS).
If you don't pray in my school, I won't think in your church.
> Unreasonable hatred is not something we wish to encourage or foster.
So why do you Red Hat shills so unreasonably hate POSIX, and the UNIX philosophy?
Unless you're running a desktop, who cares? My CentOS servers don't run Wayland, X, or any kind of graphical desktop.
They don't change major version numbers, ever. New features may be patched in as long as the risk is minimal. That's what they call "major". Maybe they should have worded it "noteworthy changes" instead.
Well, my big file server just paniced after a 6.7->6.8 upgrade. The ONLY reason to stick with CentOS6 was stability and long lifetime. Since that is now out the window, switching to Ubuntu 16.04 (with the huge advantage of having ZFS precompiled) is back on my burner.
The good news is it doesn't look like I lost any of my 24 TB of ZFS data, despite panics, reset switches, and power buttons.
No, They backport fixes, which is why the full version of the 2.6.32 kernel used in CentOS 6.8 is kernel-2.6.32-642, this means that it's has seen 642 revisions since the original release of the 2.6.32 kernel.
Such as?
So you sysv lovers now also love upstart?
I strongly suspect RedHat has a shitload of paying customers who have told them in no uncertain terms that if RH drops RHEL6 and goes systemd-only, they'll be moving to a non-systemd distro.
I can assure you, systemd would be far, FAR down the list of reasons for RHEL's paying customers to not switch from 6 to 7.
You misunderstand badly, in fact very badly, and it's not even about whether systemd is any good or not.
Workstation software development, especially the commercial closed source kind, is very slow.
A major change in how things work such as systemd means that it will be several years before the devlopers of that sort of software even consider getting their software to work in the new environment instead of just telling users to use the old one. See also how many commercial packages still recommend turning SElinux off because they haven't got around to working out how their software can operate in that environment.
So until a lot of vendors dust off their rc scripts and work out what they have to change to run on a host with systemd there are plenty of their clients who will not be upgrading due to systemd itself.