On 03/31/2015 09:53 AM, Ryan Qian wrote:

As a CentOs newbie, I'm not sure, will we still have CentOS 7.1 which derive from RHEL 7.1? or this is the new naming conversion for CentOS 7.

Thanks! -Ryan

That was going to be my question as well. According to http://lists.centos.org/pipermail/centos-announce/2014-July/020393.html the convention (for the 7.0 release at least) says:

"Numbering

CentOS 7.0-1406 introduces a new numbering scheme that we want to further develop into the life of CentOS-7. The 0 component maps to the upstream realease, whose code this release is built from. The 1406 component indicates the monthstamp of the code included in the release ( in this case, June 2014 ). By using a monthstamp we are able to respin and reissue updated media for things like container and cloud images, that are regularly refreshed, while still retaining a connection to the base distro version."

I would have assumed that this release would be "7.1.1503", and the URL on at least one mirror has:

http://mirror.fdcservers.net/centos/7.1.1503/

Guess if that's the new convention, I'll need to keep my ISO files sorted out somehow, as this progression isn't intuitive:

CentOS-7.0-1406-x86_64-DVD.iso CentOS-7-x86_64-DVD-1503.iso

-Greg

On 03/31/2015 01:28 PM, Johnny Hughes wrote:

On 03/31/2015 12:31 PM, Greg Bailey wrote:

...

On 03/31/2015 09:53 AM, Ryan Qian wrote:

...

As a CentOs newbie, I'm not sure, will we still have CentOS 7.1 which derive from RHEL 7.1? or this is the new naming conversion for CentOS 7.

Thanks! -Ryan

That was going to be my question as well. According to http://lists.centos.org/pipermail/centos-announce/2014-July/020393.html the convention (for the 7.0 release at least) says:

"Numbering

CentOS 7.0-1406 introduces a new numbering scheme that we want to further develop into the life of CentOS-7. The 0 component maps to the upstream realease, whose code this release is built from. The 1406 component indicates the monthstamp of the code included in the release ( in this case, June 2014 ). By using a monthstamp we are able to respin and reissue updated media for things like container and cloud images, that are regularly refreshed, while still retaining a connection to the base distro version."

I would have assumed that this release would be "7.1.1503", and the URL on at least one mirror has:

http://mirror.fdcservers.net/centos/7.1.1503/

Guess if that's the new convention, I'll need to keep my ISO files sorted out somehow, as this progression isn't intuitive:

CentOS-7.0-1406-x86_64-DVD.iso CentOS-7-x86_64-DVD-1503.iso

Please take a look at the "Archived Versions", and the Release Announcement:

They both tell you that 7 (1503) is derived from Red Hat Enterprise Linux 7.1 Sources. So, yes, this release, that you quoted in the Subject, is indeed exactly what you said.

And yes, this is how we are now numbering CentOS releases for 7 and greater.

OOPS: Archived Versions, on this Page:

http://wiki.centos.org/Download

On Tue, 2015-03-31 at 13:28 -0500, Johnny Hughes wrote:

On 03/31/2015 12:31 PM, Greg Bailey wrote:

...

CentOS-7.0-1406-x86_64-DVD.iso CentOS-7-x86_64-DVD-1503.iso

Please take a look at the "Archived Versions", and the Release Announcement:

They both tell you that 7 (1503) is derived from Red Hat Enterprise Linux 7.1 Sources. So, yes, this release, that you quoted in the Subject, is indeed exactly what you said.

And yes, this is how we are now numbering CentOS releases for 7 and greater.

Isn't that illogical ?

If there is:-

CentOS-7.0-1406-x86_64-DVD.iso

then the next one should logically be named:-

CentOS-7-1503-x86_64-DVD.iso

assuming sub-version numbers have been abolished by Centos.

Jumbled confusion, like CentOS-7-x86_64-DVD-1503.iso, is messy and illogical.

What is preventing Centos adopting a simple, neat, tidy, sensible and logical approach ? For example:

{major version}-{build number}-{architecture}-{media}.iso ?

That is method I would use.

Thank you.

On 03/31/2015 05:56 PM, Always Learning wrote:

On Tue, 2015-03-31 at 13:28 -0500, Johnny Hughes wrote:

...

On 03/31/2015 12:31 PM, Greg Bailey wrote:

...

CentOS-7.0-1406-x86_64-DVD.iso CentOS-7-x86_64-DVD-1503.iso

...

Please take a look at the "Archived Versions", and the Release Announcement:

They both tell you that 7 (1503) is derived from Red Hat Enterprise Linux 7.1 Sources. So, yes, this release, that you quoted in the Subject, is indeed exactly what you said.

And yes, this is how we are now numbering CentOS releases for 7 and greater.

Isn't that illogical ?

If there is:-

CentOS-7.0-1406-x86_64-DVD.iso

then the next one should logically be named:-

CentOS-7-1503-x86_64-DVD.iso

assuming sub-version numbers have been abolished by Centos.

Jumbled confusion, like CentOS-7-x86_64-DVD-1503.iso, is messy and illogical.

What is preventing Centos adopting a simple, neat, tidy, sensible and logical approach ? For example:

{major version}-{build number}-{architecture}-{media}.iso ?

That is method I would use.

Thank you.

This was discussed on the CentOS-Devel mailing list and approved by the CentOS Board. It is what we are using in the future. I suggest you become familiar with it.

Le 31/03/2015 23:24, Alain Péan a écrit :

It seems that also the redhat-release file has changed.Previously, it was : [root@centos7 ~]# cat /etc/redhat-release CentOS Linux release 7.0.1406 (Core)

Now it is : [root@centos-test ~]# cat /etc/redhat-release Derived from Red Hat Enterprise Linux 7.1 (Source)

It is also my opinion that the name CentOS-7-x86_64-DVD-1503.iso is rather confusing, it is not immediately evident that it is release 7.1. I would have prefered the name CentOS-7.1-1503-x86_64-DVD.iso, following the previous name convention.

After Karanbir answer, the redhat-release file has indeed changed after a new 'yum update'. It it now : [root@centos-test ~]# cat /etc/redhat-release CentOS Linux release 7.1.1503 (Core)

Thanks. It could indeed impact such tools as Dell OMSA ant a lot others I think.

Alain

This was discussed on the CentOS-Devel mailing list and approved by the CentOS Board. It is what we are using in the future. I suggest you become familiar with it.

Obviously naming conventions should provide for an easy upstream vendor version reference?

On 04/01/2015 11:45 AM, Александр Кириллов wrote:

...

This was discussed on the CentOS-Devel mailing list and approved by the CentOS Board. It is what we are using in the future. I suggest you become familiar with it.

Obviously naming conventions should provide for an easy upstream vendor version reference?

does /etc/centos-release-upstream provide you with that ?

On Wed, Apr 1, 2015 at 8:23 AM, Александр Кириллов nevis2us@infoline.su wrote:

Karanbir Singh писал 2015-04-01 14:25:

...

On 04/01/2015 11:45 AM, Александр Кириллов wrote:

...

This was discussed on the CentOS-Devel mailing list and approved by the

...

CentOS Board. It is what we are using in the future. I suggest you become familiar with it.

Obviously naming conventions should provide for an easy upstream vendor version reference?

does /etc/centos-release-upstream provide you with that ?

There's nothing of the sort in 7.0.1406. Ideally I'd like to see 7.1 in each and every rpm or iso name related to the point release. I'm not going to flame over something done and buried but sometimes the decisions made by rational people are just stunningly surprising.

---

CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos

It's not surprising, it's stunningly annoying.

Those of us who manage large installations of CentOS aren't involved in the development list or the board (we don't have time).

I urge the CentOS board to reconsider such a large departure from upstream. And I urge them to reach out far beyond the devel-list for opinions as that is a distinct, and quite separate, base of thought.

And, it's not just a matter of "calling it 7.1, or whatever you like." We have many scripts and operations based on determining the "version number" and if it is inconsistent with RHEL, and logic for that matter, it is more work for those who don't need it.

Yes, I'm whining. I get that. But I think I'm not alone.

On 03/31/2015 11:11 PM, Peter wrote:

Can you please point me to the centos-devel thread that discussed changing the iso naming convention from CentOS-7.1-1503-x86_64-DVD.iso to CentOS-7-x86_64-DVD-1503.iso? I must have missed it because I saw no mention of this change until today.

The first thread along these lines starts at http://lists.centos.org/pipermail/centos-devel/2014-June/010444.html

It is a long thread, as you should know, since you participated in it.

The key post in the thread, in my opinion, is at http://lists.centos.org/pipermail/centos-devel/2014-June/010944.html

My takeaway is that the ISO name for 7.1406 was an aberration, and that this is the new paradigm going forward. But I'll also reserve the right to be wrong.

On 04/02/2015 03:29 AM, Lamar Owen wrote:

On 03/31/2015 11:11 PM, Peter wrote:

...

Can you please point me to the centos-devel thread that discussed changing the iso naming convention from CentOS-7.1-1503-x86_64-DVD.iso to CentOS-7-x86_64-DVD-1503.iso? I must have missed it because I saw no mention of this change until today.

The first thread along these lines starts at http://lists.centos.org/pipermail/centos-devel/2014-June/010444.html

It is a long thread, as you should know, since you participated in it.

Yes I did, which is why I find it strange that making this particular change to the ISO name format is considered to have come from that thread. I don't recall seeing that exact change discussed, but I coudl be wrong, it was a long thread and I probably didn't read the whole thing in detail.

The key post in the thread, in my opinion, is at http://lists.centos.org/pipermail/centos-devel/2014-June/010944.html

I still don't see anything in that post about changing the iso name as mentioned above. Do feel free to point out specifics to me.

My takeaway is that the ISO name for 7.1406 was an aberration, and that this is the new paradigm going forward. But I'll also reserve the right to be wrong.

My point is that there was a claim by the board that this particular change was discussed extensively on the -devel list. If it was then it should be quite easy to point out the post(s) in the archives where this particular discussion tool place.

Peter

On Wed, Apr 1, 2015 at 2:23 PM, Peter peter@pajamian.dhs.org wrote:

...

My point is that there was a claim by the board that this particular change was discussed extensively on the -devel list. If it was then it should be quite easy to point out the post(s) in the archives where this particular discussion tool place.

The addition of a date reference makes sense to allow and identify respins within the life of a minor rev, but...

There were alternatives proposed, like: http://lists.centos.org/pipermail/centos-devel/2014-June/010940.html but I can't see any 'discussion' about why the weird concept of using the minor .0 in the initial iso name but dropping it out of subsequent versions was better or chosen.

I see the directory created on vault.centos.org is surprisingly sane, though, retaining the useful minor rev number.

Le 01/04/2015 22:15, Lamar Owen a écrit :

So, in essence you're saying that the builders of the OS that you use and trust for daily tasks are unwise, right? Sounds to me like you might want to use something different.

...

just the change will satisfy everyone.

It is impossible to satisfy everyone.

So, you refuse to hear your users, who have stated good arguments, for something that is not very difficult to change, the name of the iso, which is not coherent with the 7.0 name and confusing ? Yes, not very wise... Karanbir corrected very quickly the content of the redhat-release file, because it was totally different from 7.0, and broke a lot of scripts and applications.

Alain

On Wed, 2015-04-01 at 16:15 -0400, Lamar Owen wrote:

On 04/01/2015 03:33 PM, Always Learning wrote:

...

If someone (currently anonymous) at Centos says abandon sub-version numbers and introduce an illogical ISOs naming structure, a wise person will ignore that command.

So, in essence you're saying that the builders of the OS that you use and trust for daily tasks are unwise, right? Sounds to me like you might want to use something different.

No I am not as can be conspicuously seen in what I wrote. Lamar your introduction of non-relevant matters can not detract from the essential point I made:-

(1) removing sub-version numbers is wrong; and

(2) changing the ISO naming structure from {major version}-{sub-version}-{build number}-{architecture}-{media}.iso is an illogical unwise change because anyone looking at

{major version}-{sub-version}

instantly knows, for example, that is Centos 7.1 whereas

CentOS-7-1503-x86_64-DVD.iso

is baffling and one is then required to build and maintain a translation table to convert '1503' into Centos 7.1. That is frankly bonkers.

Creating confusion where there was originally none is essentially silly.

How many times has Johnny and others asserted that Centos is the same as RHEL ? More puzzling is the complete absence of logic for this detrimental removal of the sub-version number.

It is impossible to satisfy everyone.

I do not remember reading on this list any criticisms of the former, now abandoned, practise of using:-

{major version}-{sub-version}-{build number}-{architecture}-{media}.iso

On Wed, April 1, 2015 6:58 pm, Always Learning wrote:

On Wed, 2015-04-01 at 16:15 -0400, Lamar Owen wrote:

...

On 04/01/2015 03:33 PM, Always Learning wrote:

...

If someone (currently anonymous) at Centos says abandon sub-version numbers and introduce an illogical ISOs naming structure, a wise

person

...

will ignore that command.

So, in essence you're saying that the builders of the OS that you use and trust for daily tasks are unwise, right? Sounds to me like you might want to use something different.

No I am not as can be conspicuously seen in what I wrote. Lamar your introduction of non-relevant matters can not detract from the essential point I made:-

(1) removing sub-version numbers is wrong; and

(2) changing the ISO naming structure from {major version}-{sub-version}-{build number}-{architecture}-{media}.iso is an illogical unwise change because anyone looking at

{major version}-{sub-version}

instantly knows, for example, that is Centos 7.1 whereas

CentOS-7-1503-x86_64-DVD.iso

is baffling and one is then required to build and maintain a translation table to convert '1503' into Centos 7.1. That is frankly bonkers.

Creating confusion where there was originally none is essentially silly.

I agree with all of this. I'm more neutral to these changes merely because I don't rely as much on Linux as I did in the past. Still making change where there is no need for one is a bad practice. Changing of naming structure from self explanatory to obscure is not clever either.

Here are examples of well known ones who do these ("wrong") things:

1. Microsoft: often re-shuffles names and locations of yet the same tools (making justifiable new Administrator certifications, and making Windows admins look smart as they learn by heart stupid things like new locations of tools...)

2. Processor chip manufacturers with their chip notations (AMD was the first one who got me annoyed, even though I like them more than Intel)

Somebody, continue the list.

Does everybody think that CentOS with this change joins a good company (as I said I don't care much, I'll survive ;-)

Valeri

++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++

On 04/01/2015 08:12 PM, Always Learning wrote:

1. What is the logically reason for this alleged "improvement" ?

I never said it was an improvement. I just said that I didn't think it was that big of a deal, and it boggles my mind that people are calling a change of an ISO's file name 'unwise' and even comparing it to a Microsoft move. I just don't see it as being that big of a problem. Nor do I see it as an improvement. But the question was asked about where such a change might have been discussed, and I pointed to the long and drawn out centos-devel thread in which the background for the date-based numbering was beaten to death (and beyond). The CentOS devs have stated that the CentOS Board voted on it, and they have the decision-making power to do so. And they are all reasonable people.

2. How are users of all types, from all around the world, benefiting

from this change ?

This change makes it unequivocally clear that CentOS 7.1503 is not exactly the same as upstream RHEL 7.1, although it is functionally equivalent (where the meaning of functionally equivalent has been hashed to death, too, but it basically means binary-compatible but not necessarily binary-identical). Whether you consider that a benefit or not is up to you. I'm personally neutral on the issue.

Consolidating two replies: On 04/01/2015 07:58 PM, Always Learning wrote:

On Wed, 2015-04-01 at 16:15 -0400, Lamar Owen wrote:

...

On 04/01/2015 03:33 PM, Always Learning wrote: (1) removing sub-version numbers is wrong; and ...

That is a matter of opinion. In my opinion, assigning sub-version numbers to what was originally intended to be, by Red Hat, quarterly updates (almost Service Packs, if you will, much like SGI's numbering of their Foundation and ProPack products for the Altix server line) is what is illogical. Of course, the updates aren't quarterly any more, and other aspects of the versioning have morphed and changed over the years since the RHAS days (well, even back in certain branches of RHL 6.2, for that matter).

So you could read '7.1' as 'version 7 service pack 1.' My opinion is that sub-version numbers give a mistaken impression that the update number is a real 'version' when it was not originally so. Further, in reality the update number is meaningless for compatibility checks, as it is more than possible to have a fully updated CentOS x system that claims to be x.0 but has all the packages, save centos-release, of the latest x.y; further, it is easily possible to install the CentOS x.6 centos-release package on a completely unpatched x.0 system, making the contents of andy of the /etc/*-release files not terribly useful for strict versioning.

It is my opinion, although it's not a vehement opinion, that beginning the x.y practice is what was illogical. But it was done, and it is over, and I have more important things to do than gripe over semantics such as that.

Creating confusion where there was originally none is essentially silly.

I am not so easily confused by the new numbering; what the ISO is named is orthogonal to what it contains, at least in my mind.

How many times has Johnny and others asserted that Centos is the same as RHEL ?

The assertion is that CentOS is functionally equivalent to the upstream product. It is not 'the same as' nor can it be and still remove the trademarked branding of the upstream release. It is binary compatible without being binary identical. And as the meaning of 'binary compatible' has also been hashed to death, I'll not further clutter the traffic on this list about what it means. It's easy enough to read the centos-devel archives to see for yourself.

On 04/02/2015 01:12 AM, Les Mikesell wrote:

Adding the date component means CentOS may release more than one iso per RH's minor versions.

Newsflash: they already are, just not in the main releases trees. Look in http://buildlogs.centos.org/rolling/7/isos/x86_64/

I previously used the 20150228 CentOS 7 rolling Everything ISO to do a reinstall; worked great. Nice to not have to grab hundreds of MB of updates right out of the box. This was on the CentOS-Announce list, incidentally.

There isn't much of a consistent relationship between the RH release and the subsequent Centos release other than 'sometime later when it is ready'. So, given a set of Centos isos or even just the most recent, how would you know which RH release it is based on?

Hmm, maybe the name of the directory it is in and the link in the release notes? I also notice that the rolling point in time images have the full four digit year as well as month and day, whereas the 'functionally equivalent to a particular Red Hat update release' image has a two digit year, the month, but no day.

On Thu, 2015-04-02 at 00:51 -0400, Lamar Owen wrote:

Nor do I see it as an improvement.

Thank you for your considered response. If it is not an improvement, then there is no reason for the change, is there ?

In my opinion, assigning sub-version numbers to what was originally intended to be, by Red Hat, quarterly updates (almost Service Packs, if you will, much like SGI's numbering of their Foundation and ProPack products for the Altix server line) is what is illogical. Of course, the updates aren't quarterly any more, and other aspects of the versioning have morphed and changed over the years since the RHAS days (well, even back in certain branches of RHL 6.2, for that matter).

Whatever the original cause introducing sub-version numbering, that usage has become a clear progressive indicator of collections of updates within the major version.

in reality the update number is meaningless for compatibility checks, as it is more than possible to have a fully updated CentOS x system that claims to be x.0 but has all the packages, save centos-release, of the latest x.y; further, it is easily possible to install the CentOS x.6 centos-release package on a completely unpatched x.0 system, making the contents of andy of the /etc/*-release files not terribly useful for strict versioning.

I image the vast majority of Centos users will not risk doing non-standard updates on their production systems so your above concern is unlikely to occur.

...

Creating confusion where there was originally none is essentially silly.

I am not so easily confused by the new numbering;

I can not look at something labelled Centos 7.2169 and instantly know if it is Centos 7.1, 7.5 or even Centos 7.10. What's the latest version of Centos 6 ? Is it 6.32167 or 6.32782 or 6.32783 or should I be typing 6.23783 instead ? Confusion is not clarity.

...

How many times has Johnny and others asserted that Centos is the same as RHEL ?

The assertion is that CentOS is functionally equivalent to the upstream product.

If Centos is "functionally equivalent" to RHEL then common sense must dictate that the sub-version numbers should be compatible too.

On Wed, Apr 1, 2015 at 10:12 PM, Les Mikesell lesmikesell@gmail.com wrote:

Adding the date component means CentOS may release more than one iso per RH's minor versions. There isn't much of a consistent relationship between the RH release and the subsequent Centos release other than 'sometime later when it is ready'. So, given a set of Centos isos or even just the most recent, how would you know which RH release it is based on? Download, install, and read the /etc/os-release file before finding out? Or look up some other source of the missing information?

This could be that "some other source":

http://wiki.centos.org/Download (go to the "Archived Versions" section)

Akemi

<thought> minor version vs rolling Everything - this implies that the origin base has one root - but thats not the case (pragmatically speaking). There is still a "rebase" or at least a incompatibility between the minor versions (as shown recently 7.0->7.1 for kernel modules) and stated in the list by Johnny (different context, for EL6):

<quote>"As far as what kernel is designed for which release ... every point release (minor version) will have its own kernel branch associated with it..."</quote>

There exists more examples where this "borders" are important. So, breaking the borders between the minor releases will make the things more intransparent. Rolling it within a minor release is still great. </thought>

-- LF

On Thu, Apr 2, 2015 at 12:54 AM, John R Pierce pierce@hogranch.com wrote:

you guys sure get your panties in a bunch over something as silly as the iso file name.

if you don't like the name, rename it... sheesh.

I'm not bothered so much by the actual name as by the justification of it having been discussed on the -devel list - where in fact pretty much all of the discussion was that the minor rev number was important and should stay in.

On Wed, 2015-04-01 at 22:54 -0700, John R Pierce wrote:

you guys sure get your panties in a bunch over something as silly as the iso file name.

You may wear them, many of us don't :-)

if you don't like the name, rename it... sheesh.

Its about a consistent and logical approach to identifying versions, revisions and differences between changes.

How is the latest numbering system an improvement ? Marks idea of

{major version}-{sub version}-{mmdd} ......

is clearly a good proposal

On 04/02/2015 10:59 AM, Phelps, Matthew wrote:

It's not just the name of the ISO file. c.f. the VERSION_ID variable in /etc/os-release

In that particular place it is actually rather important, but that is orthogonal to the ISO name.

On 04/01/2015 10:10 PM, Lamar Owen wrote:

...

Yes, not very wise... Karanbir corrected very quickly the content of the redhat-release file, because it was totally different from 7.0, and broke a lot of scripts and applications.

The issue of the content of redhat-release was a serious and valid one that actually broke stuff; the ISO name being different from expected doesn't break stuff. If the ISO name broke stuff, then that would be different, and it would have already been fixed.

This is the key bit here - were not trying to break anything - we are trying to keep things sane for folks who are already invested in the platform while also allowing other people to do interesting things.

You brought up the rolling builds, and yes - we've been doing it. Its been immense value ( over 100K downloads for the 2015 Feb builds as a point of reference ). for other process's like the Atomic builds, needing the nightly, weekly builds has been key to their ability to get the technology moving. There are a lot of such examples around, but I will admit many of them are silo'd away slightly from this list - eg. the docker traction we have does not typically feedback here to this list and perhaps not into this audience. But should you want to move into container space, CentOS today represents one of the best all around experiences either on host or instance side.

The story is similar on the Cloud side of things, we have a ver large cloud instance base - if i were to fancy a guess, I'd say 20% of the centos install base is today in cloud ( and this is largely only offprem, including onprem the numbers might go higher, but there is no way to tell since they could just be real machine instances ).

Putting all that aside, the baseline assumption that we should all be mindful of is that in real terms there is no CentOS point release : any centos install, regardless of where it originated from, yum updated to the same point in time will have the same package manifest, and should deliver the same feature set ( with some exceptions, like environ specific workloads might have local flavour - you wouldnt expect cups on an GCE instance for example ).

So what the changesets have done ( and lets be fair, these changes like the iso name changing to line up a date stream rather an an arbitary point in time isnt a huge deal, you can just rename the file locally if you so wish ), effectively line up and help deliver on that.

Let me highlight this with two examples: - the Amazon instances we provide have been updated out of band, to cover for the security issues that have wider impact, heartbleed and poodle and all that stuff : labelling those as 7.0 makes no sense since it does not deliver on a 7.0 feature set, it delivers a CentOS-7/Dec 2014 update set.

- Second example is that when you look at a machine and it says 5.5 its hard to explain to the user that his machine might be 5 years out of date, there is a baseline cultural expectation that a release is either maintained or not - and having the conversations around CentOS-5 being maintained but not 5.5 isnt easy. Remember that this list represents the folks who really know, and know well, both the ecosystem and platform as well as their workloads and userbase - there are a lot ( a majority ? ) of CentOS user who dont get this. For them to line up with a CentOS-7/2014-06 and CentOS-7/2015-03 immediately makes sense. it makes it easier for us to communicate the delta in security and bugfix that they dont have on there.

And I think the overall solution we have in place right now, really does this well in that we clearly communicate the upstream relationship, while still being able to deliver the common message on and around the centos-release spec. If there are tangiable situations where this change causes harm, then I am very willing to reach out and help find a solution : dont want to break existing installs nor reduce the info available.

The other thing here in this conversation is also that there is a large emotional resistance to change. Folks expect the numbers to line up in a specific manner, and they dont - the contents of the images however still give you the metadata you need ( both ways, they should give you point in time, and release from rhel that we derived/built from ).

Happy to pickup and work through individual concerns people might have around this. But in real terms, please note that there is no change in the content being delivered. We are only opening up options to line up various media and point-in-time images.

On Thu, April 2, 2015 9:52 am, Always Learning wrote:

On Wed, 2015-04-01 at 22:54 -0700, John R Pierce wrote:

...

you guys sure get your panties in a bunch over something as silly as

the

...

iso file name.

You may wear them, many of us don't :-)

...

if you don't like the name, rename it... sheesh.

Its about a consistent and logical approach to identifying versions,

revisions and differences between changes.

How is the latest numbering system an improvement ? Marks idea of

{major version}-{sub version}-{mmdd} ......

is clearly a good proposal

After all I decide to add "<rant>" tag at the very beginning of my message instead of just assuming it. Bu before that:

Thanks a lot to CentOS team for the great job you guys are doing!

<rant>

My guess is the lack of understanding of (and sympathy to) your, Mr. Always Learning, point stems from people missing the very basic thing. I'll try to explain what I mean.

Us, human, usually do consecutive counting as follows:

A:

0 1 2 3 4 5 6 ...

Now, as portion of version identifier doesn't follow this way of counting anymore, it is akin counting like:

B:

231 2735 2746 3458 5216 ...

This is still in ascending order, still:

1. whereas in case A given you have [sub]version number 4 you definitely know that adjacent previous is 3 and adjacent following will be 5. Case B is different: unless you have the whole row of legal numbers in front of you, you will not be able to guess whether 2746 and 3458 are consecutive versions, or there is one or more versions between them.

2. comparison of two version in case A easily reveals which is earlier and which is higher, in case B it is not quite so (you can try to time yourself on comparison of random natural number in 10000 range and compare that to the case of natural numbers 0-9, you will know what I mean), and hence prone to higher chance of error (and don't second guess me: I always has A+ in mathematics in school and university ;-). This is just a trivial human psychology...

Valeri

PS I do realize that these big numbers are quite likely just a subset of indeed consecutive natural numbers, say, counting builds, and only the ones that are good enough to be released for public use are visible to public. Still, developers usually have their magic way to keep track of their consecutive builds and relation to still consecutively numbers "good" build released to public. Abandoning that is not wise at the very least. It converts product from being transparent to getting obscure for everybody. Which only serves the goal of diverting people to much poorer IMHO alternatives, MS Windows to name one (the only OS of many I know whose vendor tells you it is unsafe to use it without 3rd party software - antivirus).

</rant>

You should guess all I say is ran, so I decided to drop resemblig tag at the beginning ;-)

++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++

On Thu, Apr 2, 2015 at 10:56 AM, Karanbir Singh mail-lists@karan.org wrote:

os-release has been at /7/ since the first CentOS 7 release - what extra value does having 7.1 in there bring ? At best it just says that your centos-release rpm has not been updated and/or there is no system level state change that required metadata in that file.

If you know that some feature was added or bug fixed in RH 7.1, or more relevant, your boss or security officer or application developer knows that, there is very much value in being able to say that CentOS 7.1-whatever includes the same features/fixes, and that your automated inventory database will show which machines have been updated to that version. Otherwise you'll spend the rest of the day discussing how fix x is done in package-revs-n1 fix y is in package-rev-n2 and how to check for it. Sometimes you need the latter detail, but mostly not, especially for the application guys.

Note that any CentOS machine, updated to the same point in time, regardless of where and how it was privisioned should give you the same functional package set. This is an important thing.

Yes, but how do you explain that relationship to someone who only has a summary of the RH releases or where the Centos release stands compared to it. For example, what would you have said a few days ago?

Le 02/04/2015 18:22, Les Mikesell a écrit :

...

Note that any CentOS machine, updated to the same point in time,

...

regardless of where and how it was privisioned should give you the same functional package set. This is an important thing.

Yes, but how do you explain that relationship to someone who only has a summary of the RH releases or where the Centos release stands compared to it. For example, what would you have said a few days ago?

Notice that a new minor release includes new drivers for new servers, so it is important to know if you can install at all the system on your server, before any updates !

Alain

Le 02/04/2015 18:41, Johnny Hughes a écrit :

...

Notice that a new minor release includes new drivers for new servers, so

...

it is important to know if you can install at all the system on your server, before any updates !

what does that have to do with an ISO name?

If you use the iso that does not include the correct drivers for your new server, it could be impossible to install the server. And with the confusing naming, it could be difficult to know if the iso you have at hand has indeed have the drivers you need...

Alain

On 4/2/2015 9:49 AM, Les Mikesell wrote:

How, without a cross reference of some sort, do you know if a given CentOS iso will install on hardware where you know that the needed driver was added in an RH minor rev?

always use the latest one.

On Thu, Apr 02, 2015 at 11:57:23AM -0500, Les Mikesell wrote:

...

...

How, without a cross reference of some sort, do you know if a given CentOS iso will install on hardware where you know that the needed driver was added in an RH minor rev?

always use the latest one.

Which, combined with the possibility of releasing multiples per minor rev and no determinate time frame for the actual initial Centos minor release, really means nothing.

Well...

"Always use latest one" *plus* "look for the latest release announcement".

Like http://lists.centos.org/pipermail/centos-announce/2015-March/021005.html

A cross-reference doesn't really seem necessary because usually hardware enablement is additive. Either CentOS is up to the version you need, or it isn't yet.

If you really _need_ a specific minor release and want to _stay_ on it, to my knowledge, that's not something CentOS has _ever_ done anyway. You can pay for Red Hat's "EUS", or, I think Scientific Linux actually does keep the ".y" releases separate (but I'm not sure of the details as to how that's implemented).

On Thu, Apr 2, 2015 at 12:22 PM, Les Mikesell lesmikesell@gmail.com wrote:

On Thu, Apr 2, 2015 at 10:56 AM, Karanbir Singh mail-lists@karan.org wrote:

...

os-release has been at /7/ since the first CentOS 7 release - what extra value does having 7.1 in there bring ? At best it just says that your centos-release rpm has not been updated and/or there is no system level state change that required metadata in that file.

If you know that some feature was added or bug fixed in RH 7.1, or more relevant, your boss or security officer or application developer knows that, there is very much value in being able to say that CentOS 7.1-whatever includes the same features/fixes, and that your automated inventory database will show which machines have been updated to that version. Otherwise you'll spend the rest of the day discussing how fix x is done in package-revs-n1 fix y is in package-rev-n2 and how to check for it. Sometimes you need the latter detail, but mostly not, especially for the application guys.

This is the crux of the issue in my mind. The complete departure from the upstream naming conventions, weather they are "correct" or "relevant" or not, is a major change and is becoming a major hassle, maybe not from an engineering point of view, but from a practical, day-to-day one.

Change is fine, but it requires work to deal with. And most of us don't have time to deal with major changes. This is a major change from past practice for CentOS, and there are many operational implications of it that apparently haven't been considered.

On Thu, Apr 2, 2015 at 1:08 PM, Johnny Hughes johnny@centos.org wrote:

...

If you really _need_ a specific minor release and want to _stay_ on it, to my knowledge, that's not something CentOS has _ever_ done anyway. You can pay for Red Hat's "EUS", or, I think Scientific Linux actually does keep the ".y" releases separate (but I'm not sure of the details as to how that's implemented).

That last paragraph is EXACTLY the message we are trying to put out here. CentOS releases are NOT the same as EUS and have never been .. yet that seems to be what people expect. We want there to be no doubt on this issue.

But you are adding more confusion than you resolve if the designation does not indicate that a specified version is 'at least' up to the equivalent of some RH minor rev. Even for the people who might have incorrectly thought is was pinned there. And now there yet another arbitrary difference in what you need to know about one major number vs. another for the long interval they will co-exist.

Les Mikesell wrote:

On Thu, Apr 2, 2015 at 1:08 PM, Johnny Hughes johnny@centos.org wrote:

...

...

If you really _need_ a specific minor release and want to _stay_ on it, to my knowledge, that's not something CentOS has _ever_ done anyway. You can pay for Red Hat's "EUS", or, I think Scientific Linux actually does keep the ".y" releases separate (but I'm not sure of the details as to how that's implemented).

That last paragraph is EXACTLY the message we are trying to put out here. CentOS releases are NOT the same as EUS and have never been .. yet that seems to be what people expect. We want there to be no doubt on this issue.

But you are adding more confusion than you resolve if the designation does not indicate that a specified version is 'at least' up to the equivalent of some RH minor rev. Even for the people who might have incorrectly thought is was pinned there. And now there yet another arbitrary difference in what you need to know about one major number vs. another for the long interval they will co-exist.

Let me also add to Les' argument, in that there *are* point releases - when we go from x.y to x.z, there are usually on the order of 300 packages (I believe when I upgraded two servers yesterday from 7.0 to 7.1, there were 267 or so packages updated, and I think a few installed. That's *not* the same as yum update, and I get 10 or even 70 packages.

And when you have to talk to Windowsiacs, who know nothing other than version and point, it works best to tell them we're on that point, so go away, and don't bother us....

mark

On 04/02/2015 01:28 PM, Phelps, Matthew wrote:

Soliciting our feedback *before* changing everything regarding release names would have been nice.

We did.

http://lists.centos.org/pipermail/centos-devel/2015-February/012873.html

On Thu, 2015-04-02 at 11:08 -0500, Valeri Galtsev wrote:

A:

0 1 2 3 4 5 6 ...

That is called an arithmetic progression (from my school days)

B:

231 2735 2746 3458 5216 ...

does not resemble a geometric progression.

Lets have a LOGICAL numbering system. How about

Centos 6.6-1503 ........

derived from {major}{sub}{yymm}.

But what happens when 2 or more revisions occur within the same month ?

Will we have Centos 7.1504 and 7.1504a and 7.1504c or will someone decide to use 7.1505 (= May 2015) whilst still in April ?

Clarity is important in all things 'computer'.

Valeri

1. whereas in case A given you have [sub]version number 4 you definitely

know that adjacent previous is 3 and adjacent following will be 5. Case B is different: unless you have the whole row of legal numbers in front of you, you will not be able to guess whether 2746 and 3458 are consecutive versions, or there is one or more versions between them.

2. comparison of two version in case A easily reveals which is earlier and

which is higher, in case B it is not quite so (you can try to time yourself on comparison of random natural number in 10000 range and compare that to the case of natural numbers 0-9, you will know what I mean), and hence prone to higher chance of error (and don't second guess me: I always has A+ in mathematics in school and university ;-). This is just a trivial human psychology...

Maths was my favourite school subject too.

On 04/02/2015 02:29 PM, Phelps, Matthew wrote:

On Thu, Apr 2, 2015 at 3:25 PM, Jim Perrin jperrin@centos.org wrote:

...

On 04/02/2015 01:28 PM, Phelps, Matthew wrote:

...

Soliciting our feedback *before* changing everything regarding release names would have been nice.

We did.

http://lists.centos.org/pipermail/centos-devel/2015-February/012873.html

WE'RE NOT ON THE DEVEL LIST!

WHY NOT?

Yes, I'm shouting.

I DON'T KNOW WHAT WE'RE YELLING ABOUT

Seriously though, the change was brought up over a month ago on the -devel list. When we're asking for feedback about possible changes, it's on the -devel list, because it's about the development of the distro. If you want to give input for the direction of the distro, that's the place to do it. I would encourage anyone who's interested to join.

On Thu, 2015-04-02 at 13:08 -0500, Johnny Hughes wrote:

CentOS releases are NOT the same as EUS and have never been .. yet that seems to be what people expect. We want there to be no doubt on this issue.

Is there a commercial motive for this 'unwelcome by most' change ?

If Centos is the same as RHEL then RH can loose valuable sales income because customers, actual and potential, use free Centos. However if Centos version numbers are vastly different from RHEL version numbers and there is no reliable method of equating Centos sub-versions with RHEL sub-versions, RH gains extra sales because of the uncertainty of Centos being 'just like' reliable RHEL.

Just a casual thought.

On Thu, 2015-04-02 at 14:25 -0500, Jim Perrin wrote:

On 04/02/2015 01:28 PM, Phelps, Matthew wrote:

...

Soliciting our feedback *before* changing everything regarding release names would have been nice.

We did.

http://lists.centos.org/pipermail/centos-devel/2015-February/012873.html

Strange - I thought this mailing list was named Centos Users !

On 04/02/2015 03:12 PM, m.roth@5-cent.us wrote:

And when you have to talk to Windowsiacs, who know nothing other than version and point, it works best to tell them we're on that point, so go away, and don't bother us.

They know Service Packs and Build numbers. Call it CentOS 7 Build 1503 if you'd like. They will understand that nomenclature just fine.

On 4/2/2015 1:16 PM, Always Learning wrote:

Although most people in the world will privately complain the vast majority do not complain in public. Where is your contrary evidence that this non-beneficial and illogical change is welcome by the majority of Centos users ?

you're the one claiming its 'unwelcome by most', qed, the burden of proving that statement lies entirely on you.

I for one could care less what the ISO is called.

On Thu, Apr 2, 2015 at 4:12 PM, Lamar Owen lowen@pari.edu wrote:

On 04/02/2015 03:55 PM, Always Learning wrote:

...

Is there a commercial motive for this 'unwelcome by most' change ?

Do you have data to prove that it is unwelcome by most? It is unwelcome by you and a few others I've seen comment; what percentage of the list's subscribers do you suppose that might be? (It is neither welcome nor unwelcome by me, as I've said before.)

Feedback on the direction of the distribution's development is taken on the -devel list; this list is for questions about using the distribution.

See my reply earlier. The description of the centos-devel list says "this is strictly about development."

I take that to mean it is for developers. I am not a developer for CentOS. I don't know this, but I'm guessing there are many many users and admins of CentOS who are not on that list, like me. Even a "Please check out this thread for an important discussion about the future of CentOS release names." would have been appreciated.

Now it's too late.

(As I said earlier, it's not just the ISO name either).

On Thu, April 2, 2015 2:55 pm, Always Learning wrote:

On Thu, 2015-04-02 at 13:08 -0500, Johnny Hughes wrote:

...

CentOS releases are NOT the same as EUS and have never been .. yet that seems to be what people expect. We want there to be no doubt on this issue.

Is there a commercial motive for this 'unwelcome by most' change ?

If Centos is the same as RHEL then RH can loose valuable sales income because customers, actual and potential, use free Centos. However if Centos version numbers are vastly different from RHEL version numbers and there is no reliable method of equating Centos sub-versions with RHEL sub-versions, RH gains extra sales because of the uncertainty of Centos being 'just like' reliable RHEL.

Indeed. And some conspiracy theorist might add this happened after CentOS "marrying" RH (well, getting tighter relations that is) ;-)

Just a joke to put down everybody's fighting mood (or likely switching fire onto myself ;-(

Valeri

Just a casual thought.

-- Regards,

Paul. England, EU. Je suis Charlie.

---

CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos

++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++

On 02/04/15 20:47, Jim Perrin wrote:

On 04/02/2015 02:29 PM, Phelps, Matthew wrote:

...

On Thu, Apr 2, 2015 at 3:25 PM, Jim Perrin jperrin@centos.org wrote:

...

On 04/02/2015 01:28 PM, Phelps, Matthew wrote:

...

Soliciting our feedback *before* changing everything regarding release names would have been nice.

We did.

http://lists.centos.org/pipermail/centos-devel/2015-February/012873.html

WE'RE NOT ON THE DEVEL LIST!

WHY NOT?

Would it have made a difference? Yes, you asked on that list. And yes, nearly everyone who responded said no to the change, yet you did it anyway.

If Matthew had been on the list, and had also responded against the change, would his voice have made the difference?

Seriously, show me a post in that thread from someone not on the CentOS board speaking strongly in favour of the change. It's a big thread.

The point is, you asked and the community said no (at least those who were subscribed to the -devel list and took the time to respond). Yet you did it anyway.

The damage is done now, you can't take it back.

On Thu, 2 Apr 2015, Les Mikesell wrote:

Well if you define 'always' as 'for CentOS6 and later...

Yes, you are right. I was relying on my obviously faulty and aged memory, so I checked on my two remaining CentOS 5 boxes. There is no /etc/centos-release file there at all, only an /etc/redhat-release, so obviously not a symlink at all. More coffee.

Steve

On 02/04/15 21:35, Phelps, Matthew wrote:

See my reply earlier. The description of the centos-devel list says "this is strictly about development."

Matt, come join the contributor base - be a commnuity communication liason ( or, I am sure we can find a title to quantify this ).

stretching this a bit futher : lets see if we can find 10 people who might be considered 'community beacons', who could / would act as commnuity comms and liason to make sure we are driving in the right directions and communicating things in the most impactful manner.

I am willing to lobby the board to then allow this group to spectate and feedback into Board Meetings ( we meet once a month ).

One data point I want to drop in here is that less than 0.1 % of the CentOS user base has any contact with the project ( wherein I imply, lists + forums + irc + bugs + wiki ), so we might need to spread the net wide to find a reasonable representation.

thoughts ?

On Thu, 2015-04-02 at 17:37 -0400, Lamar Owen wrote:

On 04/02/2015 04:16 PM, Always Learning wrote:

...

Although most people in the world will privately complain the vast majority do not complain in public. Where is your contrary evidence that this non-beneficial and illogical change is welcome by the majority of Centos users ?

The burden of proof for your statement is on you, not me. I never said it was welcome by most, either. I might even agree with your statement, for that matter; but you made the statement; you have the burden of proof of that statement. I do not need to prove the converse.

Lamar,

You are a really great guy. Earlier you admitted the change was not beneficial and now you seem to agree the change was NOT welcome by most.

Query .... why did Centos bosses make the change ?

On 04/02/2015 07:00 PM, Karanbir Singh wrote:

stretching this a bit futher : lets see if we can find 10 people who might be considered 'community beacons', who could / would act as commnuity comms and liason to make sure we are driving in the right directions and communicating things in the most impactful manner. ... thoughts ?

I really think that if someone is actually interested in helping the project, rather than being a backseat driver and griping at every change from the Steeped Tradition of the Unix Protectors - In Training , a division of the National Organization of Whiners (STUP-IT / NOW) standards, that someone should be willing to take the initiative to follow the -devel list. (Yes, that contrived acronym is tongue-in-cheek and meant as a joke to lighten things up a bit.... no offense to any particular person intended; please take a good laugh, smile, and enjoy your Friday!)

I am as pressed for time as anyone else on this list; I especially feel Matt Phelps' pain, as part of an educational institution where funding and staffing is never enough. But there has to be a bar to meet so that feedback given is useful and not trollish.

I would suggest that a periodic informational FAQ be added to the monthly mailman reminders for the CentOS lists that can give a pointer to those who would like to give feedback, or help out, or otherwise do something to benefit the project as a whole.

I would also suggest that changes to the distribution that directly affect users and users' expectations be more widely announced, and something like a request for comment be made for the proposed change, with replies to be sent to the -devel list. While I consider the very specific issue of the ISO naming to be a tempest in a teacup, I also appreciate the fact that mine is not the only opinion.

But I believe that we would experience heavy turnover in such a go-between position as you describe.

On Fri, Apr 3, 2015 at 12:19 PM, Always Learning centos@u64.u22.net wrote:

Will Centos versions eventually become incompatible, partially or wholly, with its parent's RHEL versions ? I can understand why that would be commercially advantageous to RH.

I think it would be commercially advantageous if they did just the opposite - that is, make it so you could run exactly he same product on all of your machines and pay for support on the ones where you need support. I think that is the way Oracle is handling it. And their download approach makes it pretty clear that you are getting 7.1.