NanoBSD VGA Images Missing Kernel
-
New image 2.1 RC also would boot on my Hardware "Caswell CAD-2015"
It show, "Can't find kernel"
The full image of "pfSense-2.1-RC0-4g-i386-nanobsd_vga-20130627-2149.img.gz" got problem.
I have reverse to BETA1 version. :'(
We're aware, it's just the nanobsd_vga images that have a problem. We should have a fix shortly.
-
Oh please, not again… not more images without kernel. Why is the pfsense-tools stuff still not branched so that you do not kill other people's boxes? (I also do not understand what is important about this fancy kernel naming, which broke everything repeatedly.)
-
Oh please, not again… not more images without kernel. Why is the pfsense-tools stuff still not branched so that you do not kill other people's boxes? (I also do not understand what is important about this fancy kernel naming, which broke everything repeatedly.)
It's not "again" it's the same issue. Every other image has been fixed aside from NanoBSD+VGA.
And the changes are necessary to fix other issues in the builder code that prevent us from doing things like cross-compiling and properly tweaking the build with src.conf and similar. The end result is to be able to target more platforms in the future.
-
Would have to repeat myself. Branch the code. Stuff without kernel is not RC. It is just unusable. Probably the scripts could do with some sanity checking as well so that such images just cannot get published.
-
Would have to repeat myself. Branch the code. Stuff without kernel is not RC. It is just unusable. Probably the scripts could do with some sanity checking as well so that such images just cannot get published.
You're not alone in that thinking, but it's not that simple.
-
You're not alone in that thinking, but it's not that simple.
Seems extremely simple to me (esp. with git) to not use completely untested code for building RC images.
-
"seems".
-
Shrug. This is plain stupid, end of story. You do not mess with build scripts used to generate images that you advertise as has no significant known regressions from prior releases.. It is really that simple. Want to play with bleeding edge FreeBSD code? Fine, nothing wrong with that. Do it outside of code used by general user community. Looking at the pfsense-tools commit log, looks very much like "trial and error" method of development… again, nothing wrong with that per se, except that these totally untested experiments affect to-be-released code. Stupid.
-
Snapshots are still snapshots and are inherently unpredictable from one to the next, which is why we try to test them before we deploy them.
Yes, it's unfortunate that something happened, but at this point it's beating a dead horse. It'll be fixed, and everything will be back to normal by the end of the day.
Locking the thread to avoid more non-discussion/venting. We get it.
-
The snapshots uploading now contain a proper kernel so NanoBSD+VGA should be safe again once they upload.