Red Hat Bugzilla – Bug 146451
Unmatched disks into proposed RAID causes 'long' pause
Last modified: 2007-11-30 17:10:59 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.5)
Description of problem:
When installing a software raid (RAID1 mirrored) parition onto disks
with different current data, the installer hangs during the formatting
of the partitations.
It appears that some final part of the mke2fs stage is waiting for the
software raid to complete it's resync between the two disks. This is
being done, despite the partitions being 'new' (perhaps it would be
faster to write zeros?).
The installer should somehow detect that the hang is due to an MD
resync, and display progress or something to give the user a clue that
the kernel is really hard at work, even if it takes 2 hours.
This was with /boot on /dev/md0, and an LVM partition on /dev/md1
After that is finished, the install is perfect
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Use two disks with very different data for the install
2. Select Software raid (mirrored), installing at least 2 mirrored
partitions. Use LVM for /, on 2nd raid partition.
3. Formatting hangs at the end of the second raided partition.
Actual Results: Very long pause. Only clue if you actually 'cat
/proc/mdstat' on the console. Installer appears hung, except for the
hourglass turning over.
Expected Results: A message inidcating progress, or at least a more
I realise that expecting the kernel not to rebuild as soon as possible
is a bit much, but it would be nice if it did not block the LVM formats.
With raidtools, we didn't do the resyncing at install time, but that
has its downsides too...
Fedora Core 3 is now maintained by the Fedora Legacy project for security
updates only. If this problem is a security issue, please reopen and
reassign to the Fedora Legacy product. If it is not a security issue and
hasn't been resolved in the current FC5 updates or in the FC6 test
release, reopen and change the version to match.
I recently did an install of CentOS5, and didn't see anything like this. I
presume therefore that this has been fixed.