Comment 12 for bug 1431650

Revision history for this message
Mauricio Faria de Oliveira (mfo) wrote :

Updated replies to earlier comments:

comment #6
> I think this all goes back to multipath-tools updates that would help a whole lot; but they may be a little too intrusive past feature freeze -- I'm looking into just updating the udev rules now.

Just to clarify/respond to that (earlier than patch) comment with regards to the patch attached.

Given that the patch introduces only 2 changes, which are minimal/small (and one may sometimes be a nop), and is totally contained in multipath stuff, I think it'd be OK to get it in.
(Very respectfully,) it should not make multipath less functional than it is now. :) And can't break non-multipath stuff.

comment #8
> the installed system will come up using /dev/sdb1 (for example) rather than the equivalent device mapper device,

The patch should help* with that issue (described in bug 1429327 comment #10), because the multipath disks now may* show up before the wait-for-root call.. but a proper fix should probably follow the discussion in there. I'll try to get something addressing this too, according to Steve's suggestion.

> [...] and no swap.

This one should be fixed w/ 1430074.
The /etc/fstab file used to have the incorrect/p separator for the swap device too, so it wasn't found/timed out.