To close the loop, this problem turns out to be a cobbler/koan issue. the kernel really IS invalid. with the help of "high_roller" in #rhel, i found that koan's pulling the vmlinuz kernel from the cobbler server over an http:// connection -- rather than an apparently required https:// -- and loading the resultant error-page html INTO the vmlinuz file used to (attempt to) create the PV. i.e., the pull occurs as expected, no error about the pull is launched, but the process populates the target DL with error code. of course, this results in the 'kernel', now populated with garbage, being identified as "invalid". since i'd checked my source files, and as i read the logs to imply the pull was happening without error, I simply didn't look at what actually WAS pulled. i'll communicate with the cobbler/koan author to iron this out. hope someone else finds this helpful. jimmy -- Fedora-xen mailing list Fedora-xen@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-xen
