We have a working T30 board with a root filesystem we want to put on other boards too. So we replaced the file system in a bitbake build with that one. Then creating the image (update.sh -o) to flash over tftp.
After flashing the size of the root file system given with df is just as big as the root.ext3 image. But the partitioning was done correctly (fdisk). In the update.sh script we read that the rootfs will get resized later on the target.
How and where gets the resize done? Because doing a resize2fs on each target manually is a bit exhausting.
Of course the resizing is only done on first boot with an untouched root file system which is one of them many reasons just copying/deploying root file systems like you are trying to is a bad idea.
Thanks for the reply. I’m of the mind too that it isn’t a good idea. Therefore I’m also working on the solution with layers and receipes.
But your answer leads me to another question.
If you create an image with bitbake and after you make some changes in the rootfs like edit or copy some files (or chroot and make a link or whatever), would this mean the same for resizing?
I don’t quite understand your question but maybe looking at the recipe may answer it.
My question was:
If I do some changes in the folder rootfs (from the image i.e. Colibri_T30_LinuxImageV2.6_xx.tar.bzip2), would these affect the resize?
But with your link, you gave me the answer, that it does not affect the resize.
