Announcement

Collapse
No announcement yet.

Rocky Linux Shares How They May Continue To Obtain The RHEL Source Code

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • marlock
    replied
    oiaohm
    thanks for the piece-by-piece response! it's been enlightening

    after reading it, i'd imagine they can rescind their contracts under the argument of security fraud when this can be proven (eg: when the bug report has no RHEL logo where it should), even under current contracts...

    ...and they can probably also argue successfully in court that they're not hurting the GPL (at least if they only rescind support contracts that have been abused but keep the source code publicly available), but IANAL


    IMHO the other part of their response, where they make it harder for a 3rd-party to get a copy of their source code to prevent a clone distro from being made at all is probably still a step too far under this line of argumentation and may put them in hot waters at some point in the future, let alone burning community goodwill towards them (arguably already happening, but no idea at what scale)
    Last edited by marlock; 12 July 2023, 04:38 PM.

    Leave a comment:


  • virr
    replied
    Originally posted by grung View Post

    In my opinion it is very bad idea as there is no upgrade path for Centos stream.

    I could be wrong be I'm pretty sure that you can't migrate from stream 8 to stream 9.

    For personal use it is better to use a distro which you can upgrade - that's my view. Something like Debian ,Suse or Ubuntu.
    Totally agree, Linux Mint need a decade to make upgradeable to next version be a feature. But in production level, even I accept that using a dev version of RHEL, no migration path is unacceptable. It's not wrong when saying that CentOS Stream is not ready product, AFAIK Meta had to customize CentOS Stream to continue usage on their system.

    Leave a comment:


  • virr
    replied
    Originally posted by ppcorp View Post

    You're a bit too late. Oracle has been doing exactly that for a while
    Oracle and RH used to be partnered, even before IBM acquisition. Also Oracle didn't clone bit-to-bit like "someone else", they add their proprietary software, and they only repack bug update after a period. So you get slower bug update with OL, and in larger scale, it could be cost more than direct support from RedHat, just because Oracle are not bug fixer. The argument that Red Hat is doing all the work and Oracle just copies is not true!

    Leave a comment:


  • oiaohm
    replied
    Originally posted by marlock View Post
    1) does not change my opinion about it not being OK for RedHat to undermine the GPL
    2) the insurance metaphor doesn't map cleanly to the software support scenario because insuranced goods are not GPL-licensed
    Sorry insurance metaphor is closer than what you think because support contracts do technically fall under insurance laws.

    Originally posted by marlock View Post
    if you reproduce bugs in the single RHEL machine and ask for support for it
    Do note the If there and then wake up people have been open issues with redhat that when you look closely the redhat trade marks are missing with rocky or some other 1 to 1 clone had been used. So the person did not check before submitting issue to redhat that was issue on the single RHEL machine. Yes there are other cases where it been compiler bugs effecting the 1 to 1 clones causing the problem that RHEL did not suffer from because they did not build the binary with the same version of compiler. Yes some of the complier cases have cost redhat staff hours attempting to reproduce a bug that they were never going to reproduce with RHEL because its not a RHEL problem.

    Originally posted by marlock View Post
    sure, again, dick move... but not illega...l even if you can prove it's what's happening, you'd probably be bound by contract to fix the issueon that single machine, provide the source forthe fix, etc... so you need to devise a better business model around this somehow, or live with it
    With the contracts parties sign to have redhat support there is a good chance that it is illegal todo the dick move under fraudulent representation laws but the problem also that it going to cost redhat/ibm a lot to take a party to court over it. The problem here is that is simple to prove this dick move has been happening to redhat but its not simple/cost effective for redhat to go hey dick stop doing this stupidity and pay up.

    Now as you said Redhat has to look at their business model.
    Yes nightmare here is valid options:
    1) not release all the MIT and the like license stuff that does not require them to release the source
    2) use the written source code provide option of GPL and delay provide of source for 30 to 90 days.
    3) Make SRPM that confirm to the bare min GPL requires. Again another legal option.
    4) Rework their contracts to make sure reporting a bug about a clone to Redhat support is legal to end support contract and like demand payment of a full redhat license per staff member in company because someone in the company has been dick enough todo this.

    Rocky and others need to wake up bad things have been happening to Redhat that costing them resources/money and this is why Redhat/IBM is pissed. IBM legal department will go though the GPL and will work out the min possible way they can obey it to the letter of the GPL licnese. IBM legal department will already be considering not providing the MIT/apache and so on parts..

    Lot of ways Rocky should be equally pissed with the dick move happening to Redhat because Rocky support company CIQ is not getting contracts they should because people are using Redhat support for Rocky installs instead of taking out a contract with CIQ.

    marlock it takes two to tango. Redhat/IBM is not just pissed because they want to make more money alone. Redhat/IBM is party pissed because they have the documentation that people are causing Redhat support personal to waste their time looking for issues RHEL does not have because they are too cheap to buy the right contracts and report the problem to the right party and are also too cheap to test if the problem really effects RHEL before reporting to Redhat/IBM.

    Yes profit is a fact here but the dick move is a big prolbem. Yes the dick move harms rocky Linux and Oracle Linux as well as RHEL/Redhat/IBM. AlmaLinux does not sell support.

    We want think if the billion dollar company being the villain here but we need to accept that billion dollar company has a true bandit problem that is absolutely harming them and that this harm is not just restricted to Redhat/IBM. How more resources could have rocky Linux put into rocky Linux development if those exploiting redhat to support rocky had paid CIQ for that support we don't know.

    Of course I fear with IBM legal department in the mix with Redhat they are going react in the IBM way that is not going to be nice. Yes attempting to loop hole around the Redhat restriction will make IBM legal department look to more extreme options and the extrema options I can clearly see are not going to be nice.

    We really do need to sit down and start working out a clear policy/contract design on how to deal with parties who dick move with support companies like Redhat/IBM, CIQ and so on. Nice would be clause in support contracts of all involved you do the dick move and get caught you have to pay the support provider you wasted the time at max rate and pay support contract like for 10 years to the party you should have had the support contract with so that it too costly to-do the dick move.

    Leave a comment:


  • marlock
    replied
    thanks for the new angle!

    my only 2 issues with it:

    1) does not change my opinion about it not being OK for RedHat to undermine the GPL

    2) the insurance metaphor doesn't map cleanly to the software support scenario because insuranced goods are not GPL-licensed

    it is insurance fraud if people do what you described, but it's not legally fraud if you make a copy of GPL software and use it, nor if you reproduce bugs in the single RHEL machine and ask for support for it

    sure, again, dick move... but not illega...l even if you can prove it's what's happening, you'd probably be bound by contract to fix the issueon that single machine, provide the source forthe fix, etc... so you need to devise a better business model around this somehow, or live with it

    Leave a comment:


  • oiaohm
    replied
    Originally posted by marlock View Post
    Regardless of Alma and Rocky, IMHO it seems RedHat's business model / sales strategy relies too much on owning and selling their product instead of providing their service.
    The problem is how companies need Service Level Agreements to work.

    Originally posted by marlock View Post
    even if Alma and Rocky are indeed 100% "just freeloading" on RedHat's RHEL distro work, the solution to this must not come at the cost of legally undermining the software freedoms protected by the GPL

    people may agree that Alma and Rocky are "freeloading leeches" yet still defend their right to do so because they see the alternative as even worse
    The question you are not asking is how is Alma and Rocky being used by their users. Worst case where this could be really annoying and costly to IBM/Redhat is if people are paying for like Redhat support on 1 system then using like 20 to 30 Alma/Rocky systems then when they have a problem replicate it on the Redhat system then use Redhat support....

    See problem Redhat costed what they had to charge to make profit on supporting 1 system but in reality the person has 20-30 systems.

    Originally posted by marlock View Post
    meanwhile:
    - "we put together the codebase you're using and we know it like the back of our hands"
    - "our support contract is much more effective than the cloner's" (is it? if not, get cranking)
    - "if there's a bug we're the ones that will be able to fix it"
    - "we ARE the ones that have been fixing your bugs" (+ code contribuition data)
    - "if there are questions we'll surely know how to answer them"
    - "if there's a feature request we're uniquely well positioned to tackle it for you"
    This the problem with have 1 to 1 clones. Yes Redhat can be doing all those things. But then have the problem customer is exploiting the support payment system. 1 to 1 clone gives the same errors and bugs and normal RHEL. It customers who pay for 1 RHEL support license and run like 20 clones and use 1 support license to solve those issues. Yes RHEL costs a lot but having a paid staff to answer questions and having to pay share holders is not cheap.

    Originally posted by marlock View Post
    ‚Äčthis is still a direct response yet much less counterproductive to the ethos that makes RedHat viable than "this codebase is mine, all mine, and you shaun't publish it in its entirety" EVEN if the later is somehow perfectly legal (eg: by only publishing GPL bits but paywalling MIT bits and etc)
    Part of making Redhat viable is not having customers cheating Redhat support by having Redhat support more systems than they are in fact paying for.

    Delaying release of updates to the clones will have customers cheating Redhat support who lazy about it come to customer support with a issue that on the latest update of Redhat provided distribution should be fixed.

    I see the most likely problem is Redhat has found users who use Alma/Rocky and other who end up leaching off Redhat customer support. Redhat in the past did attempt to have a clause in their support contract that you could not use any 1 to 1 clones of RHEL if you wanted support but people complained about that as well.

    marlock before you say customers of redhat would not be doing what I am saying you only need to check out redhat support and you find many examples of problem.


    This ticket here any more like it kind of raises red flag. Yes the ubuntu stuff how to support this that fine. But then we want to use Rocky but then also pay the lowest amount possible to Redhat....

    Customers using Rocky and Alma have been doing the wrong things so abusing the support services Redhat provides.

    How it need to be is you use the RHEL clones and don't pay/use Redhat support at all so take care of your self or you pay for all RHEL systems and use no clones with redhat support. Attempting to stand in the middle and use RHEL clones and pay for Redhat support is a path to staff cheating the system by having a issue on a RHEL 1 to 1 clone and reporting it as happening on the RHEL paid for machine so screwing up Redhat costing for support so undermining Redhat profitability that results in piss off Redhat we are now dealing with.

    Like it or not Redhat is justified to pissed off. Now the problem is how Redhat should deal with this problem of customers cheating the support system. Delaying updates to clones makes those cheating the system more detectable.

    Remember the support cost per install is based on math that on average machine will have X number of issues this is like insurance. Think of a case of 6 block of houses where only 1 house pays insurance but when ever a item in any of the other 5 houses broken it was taken to the house with insurance to claim it on insurance the insurance company would not be putting up with this. Redhat having the same problem with support contracts.

    This is a serous problem how to run support company with Linux without having customers cheat support company out of the money they should be paid for their services. 1 to 1 clones make this cheating the support system really simple.

    Leave a comment:


  • marlock
    replied
    extra 2-cent:

    even if Alma and Rocky are indeed 100% "just freeloading" on RedHat's RHEL distro work, the solution to this must not come at the cost of legally undermining the software freedoms protected by the GPL

    people may agree that Alma and Rocky are "freeloading leeches" yet still defend their right to do so because they see the alternative as even worse

    undermining the GPL would be a bazooka shot on RedHat's own foot and most likely everyone else's


    meanwhile:
    - "we put together the codebase you're using and we know it like the back of our hands"
    - "our support contract is much more effective than the cloner's" (is it? if not, get cranking)
    - "if there's a bug we're the ones that will be able to fix it"
    - "we ARE the ones that have been fixing your bugs" (+ code contribuition data)
    - "if there are questions we'll surely know how to answer them"
    - "if there's a feature request we're uniquely well positioned to tackle it for you"

    this is still a direct response yet much less counterproductive to the ethos that makes RedHat viable than "this codebase is mine, all mine, and you shaun't publish it in its entirety" EVEN if the later is somehow perfectly legal (eg: by only publishing GPL bits but paywalling MIT bits and etc)

    Leave a comment:


  • virr
    replied
    Originally posted by torsionbar28 View Post
    Because that's how GPL works. I guess you're also so mad that Ubuntu is based off of Debian, yes? Why didn't Ubuntu start from scratch instead of freeloading off of Debian's efforts?
    Your assumption is completely wrong. Ubuntu was and still never be a Debian freeloading, from even 1st release on Oct 2004. Currently, Ubuntu is using 45% ingredients from Debian Sid, 15% from Debian Testing, all the rest is on their-highway-dev. And note that Debian Sid and Testing are on QA process of Debian, not finished product. There're thousands of libraries on Ubuntu doesn't compatible backward to Debian. On almost 20 years Ubuntu existed, they're in both good and bad actor. There're some Ubuntu devs also be Debian contributors. And Ubuntu is the source of dirty marketing that Debian is old and obsolete. By anyway, your comparison has nothing relate because Ubuntu is not freeloading of Debian, never bug-to-bug with Debian like Cent/Rocky with RHEL.

    Leave a comment:


  • marlock
    replied
    Originally posted by Doomer View Post
    ...
    I somewhat agree and I really wish for rocky/alma to win, they aren't even competing with rhel as people who get rhel subscription aren't doing it to get the sources but to get support and people who download rocky/alma would have never pay for a rhel subscription anyway, it's the piracy fallacy all over again and I'm 100% sure that a no-string-attached rhel iso would kill both rocky and alma.
    So much THIS

    Regardless of Alma and Rocky, IMHO it seems RedHat's business model / sales strategy relies too much on owning and selling their product instead of providing their service.

    Leave a comment:


  • mSparks
    replied
    Originally posted by sophisticles View Post

    If yes, could i not revise the code and enforce a different license?
    Original authors can licence their work in as many ways as they wish.
    What no one can do, legally, in any country, and especially the USA
    Is distribute other peoples work and/or modifications to other peoples work without a licence to do so.

    To have a license you must abide by all its terms given by the original author, and remain fully compliant with them.
    Redhat do not, they are pirates.

    They are also now irrelevent. since there is



    Which is 1:1 compatible with everything RH released so far and is not pirated from the FOSS community.

    Originally posted by grung View Post
    I had to reinstall it from CentOS7 to Debian as there was no free supported upgrade path to CentOS8.
    ‚Äč

    I'm pretty sure (although I haven't checked) the OL7 iso has a migration tool that would have let you update it to OL8 and beyond.
    since
    Last edited by mSparks; 03 July 2023, 06:47 PM.

    Leave a comment:

Working...
X