erik quanstrom ha scritto:
thanks, i'll look into this.
- erik
As i wrote in a previous mail, I tried this procedure:
(pcbox is a working Atom/fossil, using the key to install on dell1).
download usbinstamd64.bz2, load the usb key
pcbox% usbsd
pcbox% disk/kfs -f /dev/sdu0/fs
pcbox% mount /srv/kfs /n/k
pcbox% du -a /n/k > /tmp/du_before
pcbox% disk/kfscmd check # Only one kfs on pcbox
...
... 3005 free blocks
...
pcbox%
unmount the key from pcbox, use it to install dell1 ...
when the compilation stops, remove the key from dell1 and re-mount it
on pcbox
pcbox% usbsd
pcbox% disk/kfs -f /dev/sdu0/fs
pcbox% mount /srv/kfs /n/k
pcbox% du -a /n/k > /tmp/du_after
pcbox% disk/kfscmd check
...
... 0 free blocks
...
pcbox%
pcbox% wc /tmp/du_before /tmp/du_after
pcbox% diff /tmp/du_before /tmp/du_after
Looking at the diff output it seems that the compiled executables
are installed on the key instead of the target hdd.
This makes me think to a damned typo in a "bind" somewhere.
adriano
