Follow-up Comment #5, bug #38677 (project grub): In the interests of establishing whether there was any possibility of GRUB being confused by past versions of the VG's metadata, I deactivated the VG (vgchange -an), dd-ed zeros over the first 1MiB of sda1 and sdb1, and re-initialized the LVM metadata from a backup (pvcreate --restorefile, vgcfgrestore).
The problem persists unchanged. My guesswork following this exercise is that 'grub-probe -t partmap' is only looking at the device it was asked to discover the partmap type of, and then running across a codepath which prints a warning because it didn't discover all PVs in my LVM. This guesswork is mainly informed by having noticed that running the grub-probe -t partmap for any LV that is located on pv0 complains about not finding pv1, and vice versa. _______________________________________________________ Reply to this item at: <http://savannah.gnu.org/bugs/?38677> _______________________________________________ Message sent via/by Savannah http://savannah.gnu.org/ _______________________________________________ Bug-grub mailing list Bug-grub@gnu.org https://lists.gnu.org/mailman/listinfo/bug-grub