Ricardo Wurmus <rekado@elephly.net> writes:

> Hi Marius,
>
>> 'core-updates' has seen a lot of changes recently.  Some of the goodies
>> include […] glibc 2.27 […]
>>
>> Are there other things that should go in?
>
> I would really like to see a patch applied to glibc that ensures that
> the “prlimit64” syscall is not used when running on the RHEL 6 kernel
> (2.6.32).  The lack of this syscall on that kernel means that getrlimits
> fails, which makes it impossible to start the JVM.
>
> This problem appeared with the upgrade to glibc 2.26 already, and ever
> since I’ve been trying to minimize the damage for RHEL 6 systems where
> Guix is used as a package manager (such as the MDC).
>
> A work-around for glibc 2.27 that makes things work fine with the RHEL 6
> kernel would be very welcome!  I’ve started a branch “rhel6” where the
> default glibc has been bumped back to version 2.25 but I really don’t
> want it to be a long-lived branch; one of the reasons is that building
> all packages for this old glibc version (even just on x86_64) puts our
> build farms under extra stress that I would like to avoid.

Are you sure prlimit64 is the only syscall that needs to be adjusted?  A
quick grep through through the commit log reveals some other interfaces
that may need to be restored.

Here is the commit that removes the fallback code for missing prlimit64:
https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=695d7d138eda449678a1650a8b8b58181033353f

And here are similar commits I found by grepping the log for '3.2':
https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=e92030239abb4038d4f915d47021d6c037239309
https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=1721145f0341d70a6d7807b172c5eb400b508fc0
https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=9a45f54310573c190fa270e1f80d8307750305e9
https://sourceware.org/git/?p=glibc.git;a=commitdiff;h=e8f1225ca4d4afa4043c5267ae6dbe12268e2637

Since it's fairly late in the core-updates cycle, I think it would be
better to try restoring the prlimit64 fallback on the "rhel6" branch and
then revisit this during the next core-updates.

This got me thinking, perhaps it's possible to run Guix through a thin
hypervisor layer that uses the host virtualization facilities, or a Qemu
built against glibc 2.25.  This is similar to how "Docker" runs on
macOS, maybe it could be used for Guix in "hostile environments" too?