Re: Conceptual talk about distributed pbulk on exotic port

[Jonathan Perkin <jperkin%joyent.com@localhost>]

* On 2018-07-04 at 07:32 BST, Julien Savard wrote:

>    3. Still run on sparc hardware but with the help of emulated sparc
>    systems (via qemu-system-sparc) and with one or 2 amd64 vm/host for the
>    scan phase ( build_clients and scan_clients in pbulk.conf). All will mount
>    some NFS filesystem for pkgsrc.Complicated setup but it might works.

I would choose this option, but just use qemu for everything.  Spin up
as many processes as your hosts will take.  Our pbulk modifications:

  https://github.com/joyent/pkgsrc/commits/joyent/feature/pbulk/trunk

include patches such as:

  https://github.com/joyent/pkgsrc/commit/b3306098b1f5a4214ffcb13ba077273408d3d034

which allow you to use NFS as the primary backing store (as we do),
but use local storage for temporary scratch space, improving
performance and reliability.

I'm busy with our 2018Q2 release right now, but once that's done I'd
be happy to provide a working example of how to do this from scratch.

-- 
Jonathan Perkin  -  Joyent, Inc.  -  www.joyent.com