After weeks of testing, CentOS 7 has been released. For a list of RELEASE NOTES, please see the Wiki.
CentOS falls in line with Red Hat Enterprise Linux 7, released just 26 days ago. It is also worth noting that an ALPHA release of Scientific Linux 7.0 is also available for testing.
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(Score: 1, Informative) by Anonymous Coward on Tuesday July 08 2014, @11:16AM
summary ever....
(Score: 2) by nightsky30 on Tuesday July 08 2014, @11:34AM
I like the summaries short and to the point. :)
(Score: 5, Funny) by janrinok on Tuesday July 08 2014, @01:33PM
I could have rewritten the release notes, or cut and pasted the press release - but I thought that the links did the job just as well. And, for those that dislike release announcements, it is easier for them to skip the story without writing a comment damning all editors to a life of living hell while demanding that we all be fired....
(Score: 1, Insightful) by Anonymous Coward on Tuesday July 08 2014, @02:16PM
I liked it. Some people bitch about yet-another-software announcement. But I like them, and brief is fine for exactly the reasons you stated.
CentOS in particular, some people may still be thinking that CentOS is dead. They were inert for a very long while a few years back.
(Score: 2) by zafiro17 on Tuesday July 08 2014, @03:17PM
I'm trying to think what else you'd want in the summary. Have a click on the release notes and achieve enlightenment, punk.
Dad always thought laughter was the best medicine, which I guess is why several of us died of tuberculosis - Jack Handey
(Score: 1, Interesting) by Anonymous Coward on Tuesday July 08 2014, @12:46PM
I'm totally thrilled this project offers builds for my DEC Alpha!
(Score: 1, Funny) by Anonymous Coward on Tuesday July 08 2014, @02:05PM
DEC Alphas aside (an unusual but worthy use case, I might add) all I can say is "meh."
MEH, I say.
(Score: 5, Insightful) by evilviper on Tuesday July 08 2014, @02:24PM
XFS was "ported to the Linux kernel in 2001", the same time Ext3, Reiserfs, JFS, and others showed-up. Why was Ext3 or Ext4 ever even used, then, only to be demoted later for something that's always been around?
RedHat, other distros, and Linux in general, have a bad habit of unnecessarily indecisively waffling back and forth like this. The file system development in the BSDs has been dead simple... FFS/UFS for decades (always better than Ext2, not sure why Linux didn't use it), then add softupdates circa 2000, then UFS2, and recently ZFS or HAMMER.
Don't get me started on audio subsystems (sun, oss, alsa, esound, arts, pulse, etc), unnecessary graphical boot menus that screw up serial / ipmi, ever-changing initrd, the back and forth with dcop, bonobo, dbus hald, systemd, pam, udev, etc., and all the time some of them spend fighting with each other...
Hydrogen cyanide is a delicious and necessary part of the human diet.
(Score: 3, Interesting) by VLM on Tuesday July 08 2014, @04:59PM
"RedHat, other distros, and Linux in general, have a bad habit of unnecessarily indecisively waffling back and forth like this"
Why? Or more specifically, why is it bad? I suspect if you analyze it you'll find no reason at all.
All those FS are compatible with each other at the app layer so just use whichever fits your mood. Slap down the one of your choice and build your OS on it, the OS isn't going to care, that's for sure, and the users will never know the difference. It would be like including code to force sysadmins to only use Western Dregital or only Trashiba drives just to piss them off for fun, because you can. Sounds like something the systemd guys would do as a lullaby or give as a wedding gift.
From memory, assuming you want a FS contrast discussion, a decade ago, start with they're all identical. XFS is microscopically faster but occasionally would blow up and take all your data with it. I fooled around with it when it was new for mythtv type use, but if you want 1% faster it turns out to be immensely less of a PITA to just buy a faster disk or at current technological improvement rates, wait about a month (average). Or just stick in more ram for cache and gain a lot more than 1% performance. Or cough up the dough for a SSD, or ten of them in parallel, or whatever. The point is improving performance by screwing around with the FS might have been useful in '94 but in 2014 its a waste of time.
You'll hear a lot of "back in '93 ext2 couldn't be online resized so unknownfs is superior in 2014" BS that doesn't matter unless you get stuck in a time machine and sent back to '93.
The only feature I can recall between a ext? and a xfs is xfs cannot be shrunk. You can shrink a ext if you have to, and you can expand a xfs, but never shrink one. Weird but true. Every other feature proudly trumpeted either exists in the other FS or soon will.
I don't recall the ext to ext2 conversion (am I getting old?) but you've always been able to move up from ext2 to 3 by running a tune2fs command to add a journal, and theres some way to upgrade a ext3 to ext4. In all cases you're better off making a new one if possible, but it does kinda work.
There is a NIH aspect, or branding, maybe, thats a pretty dumb reason to make a technical choice but people do it all the time anyway. So XFS came from SGI if you worship at the church of SGI, or on the other end there's reiserfs and thats a whole nother story. I was surprised they had nothing on him but coincidence, and shockingly they turned out to be right, he really did do it, led the cops to her body and everything as I recall after the conviction.
So if you want something that works, install the second most recent ext series which today would be the incredibly well debugged and stable ext3 although ext4 is supposed to be OK. If you want something faster spend money on hardware instead of being a cheapskate. If you want to go all college indie band, alternative, mine is weirder than yours so mine is inherently better because its nonconformist, then fly your rebel flag proud and install one of about 20 FS that do the same thing as ext, slightly differently and with more bugs.
(Score: 4, Insightful) by evilviper on Tuesday July 08 2014, @06:46PM
Changing file systems affects users and admins directly. There are ins and outs of file system maintenance, backup, recovery, boot loaders, etc., etc. Several incompatible projects also affect code maturity and stability, as developer efforts are spread much more thinly, all over the place.
It affects all the interested 3rd parties, too. Ext2 was a good option for exchanging files with other systems, and Ext3 was compatible, but with the goalposts continuing to move, it seems it will never displace VFAT and NTFS, despite the compatibility difficulty, and patent issues. The more effort it takes to follow the moving target, the less likely anybody will.
Hydrogen cyanide is a delicious and necessary part of the human diet.
(Score: 3, Interesting) by VLM on Tuesday July 08 2014, @07:08PM
I donno man. Users simply don't do maint (what is that?) or backup (LOL), recovery means buy a new PC and complain about all the lost data and/or have steam and dropbox copy it all onto the new box, the OS installer sets up the boot loader and they never touch it again, ever. If you mean users as in people who hit my http server...
As for admins, again, what is file system maintenance I haven't done that unless you count defraging MSDOS partitions in the mid 80s, I don't backup systems or filesystems anymore, with virtualization and automated configuration, I only backup data, and have some moderately advanced systems to do it (although the NAS and virtualization boys are responsible for keeping their hardware up, I just get to spin up and down images on it). Since I don't do backups, restores amount to convince puppet to configure the new virtual image as a "X" and then run scripts to toss data on them. I can spin up something "new" as fast or faster than restoring it. Unless you mean lighting up a snapshot because of a big mistake, but that doesn't count and the NAS takes care of all that, not me. I run some bare iron like that, although its a lot slower than the production virtualized stuff. Other than screwing around with bootloaders for the sake of screwing around, which is kinda fun, I've never messed with them. When you replace a failed RAID mirror disk you get to reinstall grub, lilo in the old days, on the new drive, this is like a once a year or less thing so you google for it. Its not a serious problem.
I transfer files using TCPIP or much more rarely optical media. Last time I transferred data on magnetic media from one machine to another, or cross platform... I think they were still selling ZIP disks new, or maybe floppies? In fact I think it was on a ZIP disk? I haven't seen a new desktop or new laptop or new apple product with a floppy drive in quite some time... old stuff... old stuff.
"as developer efforts are spread much more thinly" LOL you go convince a FS dev to rm -RF his pet FS and work for another FS for the greater good of the community. Wake me when that happens for any FOSS project, ever. That just doesn't work. Best case might be pissing off a dev so much he quits development completely in that field and does some entirely different software. The only way to get a dev who wants to work on project ABC to work on project XYZ is to pay him. If he wanted to push changes to XYZ he'd already be doing it and you wouldn't be able to stop him, so he must not want to.
For code maturity and stability, just use something stable like the ext series, like 99% of installed boxes out there. Its not that hard.
(Score: 2) by evilviper on Wednesday July 09 2014, @04:07AM
File systems are still needed on optical media, solid state, etc. They don't cease to matter, just because you aren't using "magnetic media".
Optical discs are pretty uniform, with the UDF file system, but with all those SD Cards, USB thumb drives, removable hard drives, etc., it's a sad, sorry mish-mash of FAT32, EXFAT, NTFS, etc. An open source file system could have dominated, if there was a focused effort on one, instead of a plethora of competing almost-the-same file systems.
Wake up... It has worked with the BSD forever, not to mention every other operating system out there. Nowhere else but on Linux is it such a mess, with Linus allowing the kitchen-sink approach to development, and accepting anything and everything. And RedHat and other distros just jumping from one to another on a whim.
Hydrogen cyanide is a delicious and necessary part of the human diet.
(Score: 2) by VLM on Wednesday July 09 2014, @01:27PM
"They don't cease to matter, just because you aren't using "magnetic media"."
Sure they do, unless you're solely into experimenting or intentionally trying to create something that won't work on any appliance out there.
Just because you can put ZFS on a sdcard doesn't mean its a wise idea and almost no machines in the universe will ever be able to use it, no cameras, no cellphones, most computers... It is exactly like your UDF/iso9660 example for optical, I can burn a dvd disk with, say, btrfs, but that would be insane because I don't own a single DVD player appliance or any machine with a BIOS that use it in any way.
With a bit of extra work I could do something apparently useless that might none the less be highly useful in a way I don't yet understand to someone, is not an excuse to force the guy who does find it useful to stop, because its his fault that I'm not as creative as he is, especially if doing nothing to stop the guy wouldn't hurt anyone at all.
"Nowhere else but on Linux is it such a mess"
The fundamental problem is no one being able to find this "mess" other than located in the opinion of non linux devs, which by definition don't matter much to the linux community.
If it was a good idea, being open source, the linux community would have copied that good idea... so either its too hard to copy (LOL) or isn't a good idea (fairly likely) or is patented (LOL)
There is some toxicity in the argument such that traditionally for decades there has been a lot of PR along the lines of "you can't have linux on the desktop without (insert utter nonsequitor)" which has been said about such a wide variety of ridiculous things that any "appeal to the desktop" is automatically laughed at because the previous examples have all been jokes, which could be unfair although it usually isn't. Usually if there is no good reason for a bad idea, for lack of any better argument the "desktop" gets rolled out as a blunt weapon to try to bash a pet project somewhere it doesn't belong. We're getting there now with "integration" and "tablet" taking the blame for ideas which are often awful. I don't think this disclaimer paragraph applies in any way to systemd, or forced removal of harmless FS that someone doesn't like, however.
(Score: 2) by evilviper on Friday July 11 2014, @07:54AM
If you used Ext2, ANY computer could read it, including Android phones, some embedded devices, and more. If Linux developers hadn't gone nuts and spread their file system development efforts too thin, then something like Ext2/3 likely would have gotten popular with devices, before ExFat came along, and before Microsoft was licensing NTFS. All your devices might support Ext2 today, for just a small change in leadership.
And if Linux had used UFS/FFS instead of developing Ext2 in the early days, I KNOW UFS would have taken over the world by now. A much better file system, and one with legacy support in all Unix-like OSes the world over, from Irix to AIX to Solaris, without any of the limitations of Ext2/3.
And don't scoff at ZFS. It's still early days, and could very well take over the world, in time.
iso9660 is only useful on MS-DOS, having 8.3 file names, and many other strict limits. Instead, you have to master CDs with a mess of hybrid file systems, and extensions to work on more than one OS...
Rock Ridge, Joilet, HFS, etc., ALL are necessary on some systems, but don't work on others. iso9660 is a crappy mess, and you're holding it up as an example to follow? Insane.
That's just circular logic. It implies everyone, everywhere is technocratic, has no other motivations, and are all perfect at judging such things. Many things done in Linux have been pointless dead-ends. And there's no question that Linux has lagged behind the BSDs in hardware support and performance at various times, so your claims fall completely flat.
I will ignore the rest of your silly rant.
Hydrogen cyanide is a delicious and necessary part of the human diet.
(Score: 2) by frojack on Tuesday July 08 2014, @08:13PM
But, unless I miss my guess, all the old file systems are also supported, and you can choose one or several at install time. (At least its that way with Opensuse, which I tend to use more often than any Redhat derivative.)
So people upgrading stick with what they have. People experimenting choose the new. Everybody happy.
No, you are mistaken. I've always had this sig.
(Score: 2) by frojack on Tuesday July 08 2014, @08:25PM
As I recall, XFS was a lot less mature back then, and the safe route was EXT2/3/4. Remember it had just hit the GPL in 2000.
(I also seem to recall XFS back in those days was aimed toward a specific category of file sizes, and usage patterns.)
Its ADDITION, not SUBTRACTION. You can still use what ever you want.
No, you are mistaken. I've always had this sig.
(Score: 2) by ancientt on Wednesday July 09 2014, @04:34PM
You've hit the nail on the head. I looked into this with RHEL 7 release and found out that they did it because they are seeing a lot more installs that have better performance using XFS. You can still use any of a dozen other filesystems, but they switched to XFS because it is what the most customers would benefit from. The flipside to this is that it is only really better for people managing large file systems with large IO and pretty robust CPU, so it may be better for the average RHEL customer but not better for the average CentOS user.
Which comes with the territory, if you want to be treated like a customer, you have to purchase something, but if you want to get the software benefits without paying for support, you have to figure this kind of thing out for yourself.
From Redhat [redhat.com]:
The things that were hard for me to adjust to is not using /etc/init.d/whatever to manage processes, not having ifconfig or route and having to learn the equivilant ip commands and the new firewall manager. I'm really okay with the changes but it will take me some time to get used to. Of course you can probably manage to make it work like it used to if you really want, but I tend to do minimal installations and I want to use the recommended tools.
This post brought to you by Database Barbie
(Score: 2) by AnythingGoes on Wednesday July 09 2014, @01:54AM
This is supposed to be the "Bazaar" model of development - try it out and see what sticks.. unfortunately some bad stuff might get through, but it should (in theory), get fixed... if you look at the list of items added in the last 10 years or so, you can also see the following additions - luks, usb3, opengl 4, better 64bit support, kvm, the open source radeon and NVidia drivers.
The idea is that the various distributors are supposed to fit in what they want e.g. no system etc.. in order to meet what their users would most likely want, so I guess if you don't like what RedHat is doing, tell them off and move to another distribution.