Red Hat Bugzilla – Bug 977312
mdadm Software Raid 1 hangs on FC path loss
Last modified: 2013-06-24 13:16:25 EDT
Description of problem:
I wanted to try out mdadm Software Raid 1 and assigned two FC LUNs to a test Server and did a standard Setup. After successfull setup i deassigned one of the two LUNs to simulate a desaster szenario. The software raid became unusable.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Assign two fc luns fc-lun1 fc-lun2
2. create software raid 1: mdadm --create --verbose /dev/md0 --level=mirror --raid-devices=2 /dev/mapper/fc-lun1 /dev/mapper/fc-lun2
3. mkfs.ext4 /dev/md0 + mount /dev/md0 /mnt/test
4. deassign one of the luns from step1
- mdadm --detail /dev/md0 ---> does not terminate
- cat /proc/mdstat ---> does not terminate
- /mnt/test is not writable
- raid status info with the commands
mdadm --detail /dev/md0 or
- writeble /mnt/test filesystem
words cannot describe my disappointment
i will recheck my multipath setup, maybe multipath tries to hide the path loss ....
OK, sorry guys this was my fault (wrong multipath.conf), very embarrassing. Ticket can be closed.
OK, closing out. Glad to hear it works for you ;-)