NanoBSD won't build

P

Paul47

Hello all. I am trying to get nanoBSD to build just with the stock configuration. It has failed the same way in both 32 and 64-bit FreeBSD. The machine is an Intel D2500CC motherboard, and I am running off a flash drive.

It builds for quite a while, then according to the _.bw file it fails (or appears to fail) on a link step. However when I attempt the same command by hand, it works. Here is the last command:
Code: 
ln -fs /usr/obj/nanobsd.full//usr/src/tmp/lib/libzpool.so.2 /usr/obj/nanobsd.full//usr/src/tmp/usr/lib/libzpool.so
1 error
*** [libraries] Error Code 2 ***
...
(I typed the above in by hand, pardon any mistakes.)

I'm not quite sure what to look for next. The file does exist.

I notice there is no cfg file in /usr/src/tools/tools/nanobsd. I wonder why that is?

A search of this forum did not turn up much useful.
 
Now that I think of it, not being able to build NanoBSD using the standard configuration looks like a bug. I guess I will file one unless someone has any other ideas or reasons for this.
 
I just tried running the build again, and it ran to completion.

This was my 3rd attempt. In my first attempt I had installed 64-bit FreeBSD 9.1 on a hard drive and run from there. In my second attempt I had installed 32-bit FreeBSD 9.1 on a flash drive. It's just possible I still had the hard drive connected and booted from that, thinking I was booting from the 32-bit install. This time I had the hard drive disconnected, and certainly did run the 32-bit install. So, my guess now is that there is a problem with the 64-bit install. I have read here and there of problems with 64-bit FreeBSD and this board (specifically the video) but I thought the fix was in long enough ago to get into the 9.1 release. Who knows? I am happy now anyway.
 
OK, another theory bites the dust.

I made a short config file because I wanted to change a couple things, and ran the build again. It failed!

Again it failed in the "build world" step where the linking was going on, but linking another file. I grepped and found several "Error codes" in the log, the first one being "[.depend]". There were none of these in the previous build.

I am getting the impression there is a "race condition" in the script, meaning some steps are executing before previous steps have finished. Since I have a very slow storage medium, perhaps I see it more than others do.
 
You must log in or register to reply here.
Back
Top