Re: measuring swap partition speed

void <void_at_f-m.fm> wrote on
Date: Sat, 23 Dec 2023 12:36:50 UTC :

> On Fri, 22 Dec 2023, at 20:08, Mark Millard wrote:
> > I normally use -j$(sysctl -n hw.ncpu) for installworld but,
> > if I gather right, you effectively used -j1 (implicit), so
> > I'll do both styles, -j4 being appropriate for RPi4B's.
> 
> Interesting. Advice long ago was to not use -j for the installworld part
> so I've never used it at that stage.

https://cgit.freebsd.org/src/commit/?id=ad6481dddcaa reports:

author Cy Schubert <cy@FreeBSD.org> 2017-06-28 19:05:04 +0000
committer Cy Schubert <cy@FreeBSD.org> 2017-06-28 19:05:04 +0000
. . .
src-ad6481dddcaab3363fa5f76489ba8ec11d5a235e.zip
Allow parallel installworld (-j N) and poudriere installworld
(poudriere jail -c and poudriere jail -u) to proceed.


I'll note that at the time:

#define __FreeBSD_version 1200036

was in use.


But even older is . . .

FreeBSD 7.0-RELEASE was the first where the man 7 build page
reported:

       installworld	Install	everything built  by  a	 preceding  buildworld
			step  into  the	 directory  hierarchy  pointed	to  by
			make(1)	variable DESTDIR.

			If installing onto an  NFS  file  system  and  running
			make(1)	  with	 the   -j   option,   make  sure  that
			rpc.lockd(8) is	running	on  both  client  and  server.
			See rc.conf(5) on how to make it start at boot time.


So, as near as I can tell, only bugs have interfered with use of
-jN for installworld since sometime back.

It has been some time since I last ran into problems with such
-jN use for installworld ( or installkernel ).

Also, if one is installing to a (sub-)directory tree that is
not the live boot environment's tree, there are fewer potential
issues. I have such for chroot use and poudriere jail use.

===
Mark Millard
marklmi at yahoo.com