void <void_at_f-m.fm> wrote on
Date: Fri, 17 Oct 2025 09:45:27 UTC :

> Has anyone been able to build this port since 14th Oct?
> 
> Since 
> https://cgit.freebsd.org/ports/commit/?id=38cc3ba437b87e694d99f1faaf3152ab3709501a
> 
> ?
> 
> (tried in stable/14 and stable/15 poudriere on different hosts.
> In both cases when the build got to the actual port after building 
> all the prerequisites, the host rebooted after a short compile time,
> no error message as to why. I wonder if it's a problem with
> the machines, the builders, or the port, which is why I'm asking
> here).

At this point stable/15 and a non-debug main 16 are not all that
different. So I attempted builds of what I had for a ports tree
(from Oct 13) and then updating the ports tree and rebuilding
what changed ( PKG_NO_VERSION_FOR_DEPS=yes style ) based on my
normal environment and poudriere-devel use.

Neither failed.


But you give  little configuration information so I do not
know how well my attempt approximated your context:

RAM+SWAP == ??? + ??? == ???
Number of FreeBSD cpus for the configuration?

poudriere.conf :
USE_TMPFS=???
TMPFS_BLACKLIST=???
PARALLEL_JOBS=???    # (or command line control of such)
ALLOW_MAKE_JOBS=???  # (defined vs. not)
ALLOW_MAKE_JOBS_PACKAGES=???
MUTUALLY_EXCLUSIVE_BUILD_PACKAGES=???
PRIORITY_BOOST=???
other relevant possibilities?

make.conf (or command line control of such):
MAKE_JOBS_NUMBER_LIMIT=??? # or MAKE_JOBS_NUMBER=???


Details for my context . . .

# uname -apKU
FreeBSD 7950X3D-ZFS 16.0-CURRENT FreeBSD 16.0-CURRENT main-n281155-86d17239233e 
GENERIC-NODEBUG amd64 amd64 1600001 1600001

Ignoring FreeBSD-set-* :
kernels+world most recent: FreeBSD-*-16.snap20251014234723
jail          most recent: FreeBSD-*-16.snap20251017104531

The poudriere-devel jail was based on pkgbase using:
-C -x \
^FreeBSD-set-minimal-jail\$ \
^FreeBSD-set-devel\$ \
^FreeBSD-bootloader\$ \
^FreeBSD-src-sys\$ \
^FreeBSD-set-lib32\$

The boot kernel, boot world, and jail-world were all
based on installing official pkgbase materials, not
any personal builds. Technically I've never seen a
specification of what configures the world part of the
pkgbase builds for main.

RAM+SWAP: 192 GiBytes + 512 GiBytes == 704 GiBytes
32 FreeBSD cpus

My environment uses TMPFS_BLACKLIST and TMPFS_BLACKLIST_TMPDIR
to generally avoid large tmpfs usage (so such does not compete
nearly as much for RAM+SWAP) despite USE_TMPFS=all . No
explicit PARALLEL_JOBS limit. ALLOW_MAKE_JOBS enabled.
(So ALLOW_MAKE_JOBS_PACKAGES details do not matter.)
I have a MUTUALLY_EXCLUSIVE_BUILD_PACKAGES list. Also a
PRIORITY_BOOST list to influence some ordering to try to
avoid long running single process/single thread processes
from running by themselves as much. (These drive from
prior "bulk -ca" experimental estimating measurements.)

# ~/fbsd-based-on-what-commit.sh -C /usr/ports/
ac9b12ecd53f (HEAD -> main, freebsd/main, freebsd/HEAD) sysutils/stackit: 
Update 0.43.0 => 0.44.0
Author:     Robert Gogolok <[email protected]>
Commit:     Vladimir Druzenko <[email protected]>
CommitDate: 2025-10-13 23:13:16 +0000
branch: main
merge-base: ac9b12ecd53f23b456ab648c7d5eacca717f9fe0
merge-base: CommitDate: 2025-10-13 23:13:16 +0000
n721712 (--first-parent --count for merge-base)

The incremental build finished just fine:
. . .
[03:00:27] [01] [01:31:41] Finished   devel/electron37 | electron37-37.6.1: 
Success
[03:00:37] [01] [00:00:00] Building   net-im/signal-desktop | 
signal-desktop-7.72.1_2
[03:05:00] [01] [00:04:23] Finished   net-im/signal-desktop | 
signal-desktop-7.72.1_2: Success
[03:05:04] Stopping 32 builders
. . .
[03:05:06] [main-min-devel-lib32-amd64-default] [2025-10-17_14h07m11s] 
[committing] Time: 03:05:04
           Queued: 201 Inspected: 92 Ignored: 0 Built: 109 Failed: 0 Skipped: 0 
Fetched: 0 Remaining: 0
. . .

Maximums Observered figures:
Note: 0 < SwapUsed was never observed by my hacked variation of top.
50743Mi MaxObs(Act+Lndry+SwapUsed), 132624Mi MaxObs(A+Wir+L+SU), 143875Mi 
(A+W+L+SU+InAct)

So: Somewhere between 132624 MiBytes and 143875 MiBytes or
     so was sufficient RAM+SWAP, all RAM here.


Then updating the ports tree got to:

# ~/fbsd-based-on-what-commit.sh -C /usr/ports/
dcaa6a11ac55 (HEAD -> main, freebsd/main, freebsd/HEAD) multimedia/vlc: Chase 
net/liveMedia shlib bump
Author:     Jason E. Hale <[email protected]>
Commit:     Jason E. Hale <[email protected]>
CommitDate: 2025-10-18 02:00:37 +0000
branch: main
merge-base: dcaa6a11ac55381a248181b073e941507f5b50f5
merge-base: CommitDate: 2025-10-18 02:00:37 +0000
n722021 (--first-parent --count for merge-base)

An incremental build after that finished just fine:

. . .
[00:05:03] [21] [00:02:21] Finished   net-im/libsignal-node | 
libsignal-node-0.81.1: Success
[00:08:05] [19] [00:05:18] Finished   multimedia/ringrtc | ringrtc-2.59.0: 
Success
[00:08:09] [01] [00:00:00] Building   net-im/signal-desktop | 
signal-desktop-7.74.0
[00:12:24] [01] [00:04:15] Finished   net-im/signal-desktop | 
signal-desktop-7.74.0: Success
[00:12:27] Stopping 32 builders
. . .
[00:12:29] [main-min-devel-lib32-amd64-default] [2025-10-17_19h50m24s] 
[committing] Time: 00:12:28
           Queued: 143 Inspected: 130 Ignored: 0 Built: 13 Failed: 0 Skipped: 0 
Fetched: 0 Remaining: 0
. . .

Maximums Observered figures:
Note: 0 < SwapUsed was never observed by my hacked variation of top.
6133Mi MaxObs(Act+Lndry+SwapUsed), 71606Mi MaxObs(A+Wir+L+SU), 74577Mi 
(A+W+L+SU+InAct)

So: Somewhere between 71606 MiBytes and 74577 MiBytes or
     so was sufficient RAM+SWAP, all RAM here.


===
Mark Millard
marklmi at yahoo.com


Reply via email to