maybe copy gpg to /bin ? (it's mentioned in loop-AES.README, section
5)
Thanks for suggesting I go back and check that! For reasons that escape me right now, I had placed gpg in /sbin instead of /bin. Moving it to /bin enabled losetup.sh to run successfully, at least from the commandline.
The final stumbling block was that because /etc/rc.d/rc#.d/S00losetup (a link to losetup.sh) was called too late in the boot sequence, the boot still failed when it tried to fsck loop6 before it had been losetup'ed. I solved this by adding the losetup line to /etc/rc.d/rc.sysinit immediately before the fsck calls.
Of course this isn't ideal as I'll need to re-edit rc.sysinit everytime the init scripts are updated, but that's actually less of a concern to me than having plaintext passphrases on my system. So now that things are at least working, I think I'll revisit my original plan of having all the partitions losetup/mounted within the initial initrd, ideally after entry of a single passphrase.
FG
The final stumbling block was that because /etc/rc.d/rc#.d/S00losetup (a link to losetup.sh) was called too late in the boot sequence, the boot still failed when it tried to fsck loop6 before it had been losetup'ed. I solved this by adding the losetup line to /etc/rc.d/rc.sysinit immediately before the fsck calls.
Of course this isn't ideal as I'll need to re-edit rc.sysinit everytime the init scripts are updated, but that's actually less of a concern to me than having plaintext passphrases on my system. So now that things are at least working, I think I'll revisit my original plan of having all the partitions losetup/mounted within the initial initrd, ideally after entry of a single passphrase.
FG
