I sponsored multipath-tools as provided by Mauricio; some comments:

1) for find-multipaths> while sufficient; we set the default as
find_multipaths -- it makes sense for it to always be the default (from
what I recall of my analysis of this some time ago, anyway). Some
installs may not get multipath.conf; or it may be removed by an
unsuspecting admin (or by someone who installed multipath-tools on their
system and perhaps should not have, and then remove the file). It all in
all feels like the least intrusive way to deal with multipath disks on
any random system (but I'm happy to be shown I'm mistaken! :)

As for testing; I have *not* run this in a VM; I felt familiar enough
with the changes to go ahead and upload, however there *should* be some
extensive testing of various multipath scenarios again before release to
iron out any other last-minute bugs. We're *very* close to release by
now, it would be good to not do too many SRU updates to fix things.
Among other things, we'd really benefit from testing with Tore
Anderson's use case, which is different enough from the typical tests
being done on FC and ppc64 to expose some fun bugs.

Finally, all these changes indeed need to go to Debian; we'd really
benefit from reducing our delta even more, such as making sure all the
patches are shared, including find_multipaths if possible.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1565817

Title:
  multipath: LVM physical volumes on individual paths

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/multipath-tools/+bug/1565817/+subscriptions

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to