Subject: Volumes "gone" after adding Spare to
I just added a spare object to an MD region in my EVMS setup. After I saved
the changes I checked /proc/mdstat to see if the additional device was
listed, did not see it and started the EVMS administration tool again. On
startup, the tool (evmsgui) complains:
Feb 23 14:09:29 MDRaid5RegMgr: Region md/md0 : MD superblock has
working_disks=4, found 1.
Feb 23 14:09:29 MDRaid5RegMgr: Region md/md0 : MD superblock has nr_disks=4,
Feb 23 14:09:29 MDRaid5RegMgr: Region md/md0 : Invalid MD superblocks were
found in object(s) : [hdc5_bbr hde5_bbr hdg5_bbr ]. [hdc5_bbr hde5_bbr
hdg5_bbr ] will not be activated and should be removed from the region.
Feb 23 14:09:29 MDRaid5RegMgr: Region md/md0 is corrupt. Using the Fix...
function, it may be possible to bring it back to normal state.
All my EVMS volumes are "gone" from the administration tool, but I can see
that they are still mounted and seem to contain the data they should. They
are not listed under /dev/evms anymore.
Some details about my setup:
I had three BBR segments on three separate disks which where used to build the
BBR seg1-. .- LVM reg1 - EVMS vol1
BBR seg2 |- RAID5 md reg - LVM container -|- LVM reg2 - EVMS vol2
BBR seg3-. |- LVM reg3 - EVMS vol3
.- LVM reg4 - EVMS vol4
This is how I remember it, not sure I got the region/container relation right.
What should I do now to recover from this situation? What happened after
adding the spare object?
Needless to say I am currently in an *extremely* frightened state. I tried the
"Fix..." function, but did not save any changes. The md region is still
Please help me, I am really scared of losing all my data.
Thanks a lot,
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
Evms-devel mailing list
To subscribe/unsubscribe, please visit: