The files that have to be customized or overridden by adding hooks under
/etc are:

  /lib/udev/rules.d/64-md-raid.rules
  /usr/share/initramfs-tools/hooks/mdadm

so I think this is more of an MDADM issue instead of a generic
setup/config issue. But, that depends on which package actually owns
these files. (I didn't track that down).

That said, the 64-md-raid.rules file was changed by someone to fix
Ubuntu bug 1030292 to sort out some other issue and I found that I had
to partially *revert* that change to make Intel RAID work for me. So, if
we make Intel RAID work for me, then something else may break. Which is
preferable?

I'll try to get on help.ubuntu.com to write this up.

BTW, one of the nastiest issues with Intel RAID isn't getting it
started. Its getting it stopped correctly during a shut down. That was a
major PITA and I eventually had to hack the KDE (!!) shutdown scripts to
umount the file systems I have on RAID when I log out. So, I mount RAID
during KDM startup and unmount during KDM shutdown.

I couldn't find any other way issue the magic 'mdadm -Ss --wait-clean'
command required to cleanly stop the array so it wouldn't re-sync (re-
mirror) itself on the next reboot.

This is such a total hack that I wouldn't accept it into production
code. But, if there's a better way to force a command to be executed
during a shutdown and umount then I'd be happy to help trouble-shoot it.
I couldn't find a better method and I tried a number of different
approaches to changing the scripts under /etc to do this more cleanly
than hacking KDE startup and shutdown.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1210104

Title:
  Intel FakeRAID *Regression* in Kubuntu 12.04 with Linux 3.2.0-51

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1210104/+subscriptions

-- 
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to