opnsense build using not using all cpu cores

Started by none, Today at 01:47:48 PM

Previous topic - Next topic
Hi,

I build arm64 images for opnsense and this week I begun building for 26.7 family. Usually ports build takes 20 to 24 hours, but it is now over 3 days and not close to end.

Has the change to FreeBSD 15.1 demanded special config to use all cpu cores? FreeBSD can see all 8 cores:

---<<BOOT>>---
Copyright (c) 1992-2025 The FreeBSD Project.
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
        The Regents of the University of California. All rights reserved.
FreeBSD is a registered trademark of The FreeBSD Foundation.
FreeBSD 15.1-RELEASE releng/15.1-n283562-96841ea08dcf GENERIC arm64
FreeBSD clang version 19.1.7 (https://github.com/llvm/llvm-project.git llvmorg-19.1.7-0-gcd708029e0b2)
VT(efifb): resolution 640x480
real memory  = 16879779840 (16097 MB)
avail memory = 16487489536 (15723 MB)
Starting CPU 1 (100)
Starting CPU 2 (200)
Starting CPU 3 (300)
Starting CPU 4 (400)
Starting CPU 5 (500)
Starting CPU 6 (600)
Starting CPU 7 (700)
FreeBSD/SMP: Multiprocessor System Detected: 8 CPUs

but the ports build uses only one :(

If anyone have any hints on how to change it :)

thanks,

none
"We will call you Cygnus,
the God of balance you shall be."

The ports build one by one. Some ports use multiple CPUs, some don't. Not looking into changing this.

Poudriere exists if you want to use it, but it's overloaded for our use case.


Cheers,
Franco

Thanks for the answer.

My usual build time for it all was about 36h. Now its way more, and the box is the same. I am looking for the reason :)

none
"We will call you Cygnus,
the God of balance you shall be."

I found 15.x to be somewhat slower on IO in general. Noticed while installing images mostly. I'm not sure what to do about that.


Cheers,
Franco