On Aug 19, 2019, kefu chai <tchaikov(a)gmail.com> wrote:
after a second thought, i think, probably, instead of
patching SPDK to
cater the needs of older machines. a better option is to disable SPDK
when building packages for testing and for our official releases.
I have no real clue on what role or impact disabling SPDK plays in the
grand scheme of things, so I'm happy to defer to whoever does. I just
wish to keep my home ceph cluster running on those machines until I have
good reason to replace them. Unfortunately that's the last generation
of x86 hardware that can run with a Free Software BIOS, and OpenPower
isn't so much of an option for home use. I hope retaining the ability
to build ceph so that it runs on such old and wise ;-) machines is not a
major problem.
Next in my wishlist is to try to fix the issues that I'm told are
getting in the way of ceph's running on 32-bit x86. If anyone is
familiar with them and could give me a brain dump to get me started,
that would certainly be appreciated. I don't really have 32-bit
machines running ceph daemons, but I have often recommended ceph to
people who'd like to run it on SBCs connected to USB storage, so I was
quite surprised and disappointed to find out even x86 wouldn't work any
more. Plus, that messed up my uniform selection of packages on x86 and
x86-64 machines with a single meta-package with all the dependencies I
care for ;-)
--
Alexandre Oliva, freedom fighter he/him
https://FSFLA.org/blogs/lxo
Be the change, be Free! FSF Latin America board member
GNU Toolchain Engineer Free Software Evangelist
Hay que enGNUrecerse, pero sin perder la terGNUra jamás - Che GNUevara