The-going

A very interesting investigation. I'm only collecting a minimal image. And I have never encountered such a problem. Please replace this line: https://github.com/armbian/build/blob/d511679092f07fb9e09df180698ee7d6288cb36c/lib/debootstrap.sh#L863 ``` # DEBUG: print free space local...

Thanks for the feedback. I think this information is enough. I have added the following debugging code. ``` diff --git a/lib/debootstrap.sh b/lib/debootstrap.sh index bdf011ca3..32a725c45 100644 --- a/lib/debootstrap.sh +++ b/lib/debootstrap.sh @@...

Next step: ```diff diff --git a/lib/debootstrap.sh b/lib/debootstrap.sh index 32a725c45..a9fbaa956 100644 --- a/lib/debootstrap.sh +++ b/lib/debootstrap.sh @@ -624,7 +624,7 @@ PREPARE_IMAGE_SIZE if [[ $BUILD_DESKTOP == yes ]]; then local sdsize=$(bc -l

I have adjusted the debugging code to print the calculated and actual partition sizes. The results are printed: ``` == imagesize=808 == == *1.15 sdsize=932 == == stage after prepare...

Good. I'll think about it.

> 2\. Make `/var/tmp` a `tmpfs` while running `update-initramfs`. I suspect this is fairly easy and removes point 3 above (and the resulting dependency on the image contents). At the...

> I added `EXTRA_ROOTFS_MIB_SIZE` back then as a dirty hack/workaround.. You did everything right. We can't calculate all the possible options. This should be done by the user for his...

> Another approach would be to introduce some initrd caching This is a bad idea. > I really like the idea of using tmpfs for `/var/tmp` during initrd creation (also,...

The important line in this: > Thread(s) per core: 2 4 physical processors - four temperature sensors. My chip (Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz) has 5 sensors. One per...

At this very moment in time. ``` htop> ./htop --version htop 3.2.0-dev htop> sensors --version sensors version 3.5.0 with libsensors version 3.5.0 ``` ``` htop> sensors -u coretemp-isa-0000 Adapter: ISA...