Backport cgroups v2 to 8u

[Poonam Parhar]

Hello Severin, Andrew,

We, Oracle, are in the process of backporting a bunch of cgroups v2 fixes to Oracle jdk8u. We have started with backporting the following 6 issues first.

1. JDK-8230305: Cgroups v2: Container awareness
2. JDK-8231111: Cgroups v2: Rework Metrics in java.base so as to recognize unified hierarchy
3. JDK-8253714: [cgroups v2] Soft memory limit incorrectly using memory.high
4. JDK-8237479: 8230305 causes slowdebug build failure
5. JDK-8253727: [cgroups v2] Memory and swap limits reported incorrectly
6. JDK-8278951: containers/cgroup/PlainRead.java fails on Ubuntu 21.10 

I see that Jonathan Dowland has already submitted PRs to jdk8u-dev for the first 2 issues. 

I think we can collaborate our efforts here. In a few days, we will be completing our work on backporting the above fixes to Oracle jdk8u. After that, we can contribute the changes for the remaining four issues to jdk8u-dev. Please let me know your thoughts.

Thanks,
Poonam


On 9/30/22, 8:27 AM, "jdk8u-dev on behalf of Severin Gehwolf" <jdk8u-dev-retn at openjdk.org on behalf of sgehwolf at redhat.com> wrote:

    On Fri, 2022-09-30 at 16:20 +0100, Andrew Hughes wrote:
    > On 19:03 Wed 28 Sep     , Severin Gehwolf wrote:
    > 
    > snip...
    > 
    > > > 
    > > > Do you plan to target these all to one particular release, or just
    > > > when they happen to be integrated?
    > > 
    > > It'll depend on when PRs for them become available. We cannot integrate
    > > them as we go along, as most of them are interrelated or fix bugs in
    > > earlier versions. So the plan would be to integrate (the set of) them
    > > for the January 2023 or April 2023 release whichever we can make.
    > > 
    > 
    > So would it be worth doing this on a branch as we did with JFR? That
    > might be easier to setup now we're using git as well, but I don't
    > know how it would work with Skara.
    
    We didn't need one for 11u, so my hope would be that we don't need one
    for 8u.
    
    > We could also just use dependent PRs and only push everything when
    > it's all ready. That might be simpler.
    
    Yes, dependent PRs worked fine for 11. It should work for 8u as well.
    Your concern was about integration (reviews can happen at any time).
    This would be controlled by the approval labels, so we have a mechanism
    to control time of integration. That's my thinking.
    
    Thanks,
    Severin