On Tue, 2006-02-07 at 15:23 -0500, James Olson wrote: > The real problem is that in a striped array one of the drives will > seem to be bigger than it really is. In the 2.4 kernel the ataraid > module had a feature where it suppressed the partitions on the > individual drives from the system. In the 2.6 kernel dmraid doesn't > suppress them. So far I have found that > 1) I can't use any labels in fstab or it will trigger the seeking > errors problem. > 2) have to edit the hal filter also > 3) have to edit the lvm filter > 4) and now have to edit the udev filter. So, I can't speak about 2 or 4, but 1 and 3 should be fixed in the current fc5 test releases. #1 is addressed in libblkid (part of e2fsprogs), and sct has looked at it without complaint, so once we get a bit more testing I'll forward it on to the tytso to add upstream. #3 should be fixed in more recent lvm2 packages (upstream and fc5 test packages), in that lvm2 now prefers "leaf-node" device-mapper devices over other devices when looking for pv data. It'd still be really nice to simply "turn off" physical partitions when we're using the full device in a device-map. -- Peter _______________________________________________ Ataraid-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/ataraid-list
