We have a server with a 3ware RAID controller that went into read-only mode today due to an aborted ext4 journal. I checked the RAID status and found the following strange configuration:
//host> /c3/u1 show
Unit UnitType Status %RCmpl %V/I/M Port Stripe Size(GB)
------------------------------------------------------------------------
u1 RAID-5 INOPERABLE - - - 64K 11175.8
u1-0 DISK OK - - p0 - 1862.63
u1-1 DISK DEGRADED - - - - 1862.63
u1-2 DISK DEGRADED - - - - 1862.63
u1-3 DISK DEGRADED - - - - 1862.63
u1-4 DISK DEGRADED - - - - 1862.63
u1-5 DISK DEGRADED - - - - 1862.63
u1-6 DISK DEGRADED - - - - 1862.63
u1/v0 Volume - - - - - 11175.8
//host> /c3/u0 show
Unit UnitType Status %RCmpl %V/I/M Port Stripe Size(GB)
------------------------------------------------------------------------
u0 RAID-5 OK - - - 64K 11175.8
u0-0 DISK OK - - p7 - 1862.63
u0-1 DISK OK - - p1 - 1862.63
u0-2 DISK OK - - p2 - 1862.63
u0-3 DISK OK - - p3 - 1862.63
u0-4 DISK OK - - p4 - 1862.63
u0-5 DISK OK - - p5 - 1862.63
u0-6 DISK OK - - p6 - 1862.63
u0/v0 Volume - - - - - 11175.8
I also have several of the following warnings in the log:
c3 [Wed Oct 08 2014 02:08:02] WARNING Sector repair completed: port=7, LBA=0x27000380
There is also a bit older message about an incomplete array followed by a "drive inserted: port=0" message.
The p0 disk is supposed to be a spare, but suddenly it is the only disk in an array that looks identical to the main array. Any idea what happened here?
My plan to fix this would be to remove p0 from array u1, remove the entire array u1, and mark p0 as spare. After that I'd remove p7 from the array, as it seems to be unreliable and rebuild the array using p0.
Does this look like a sensible plan? I'm a bit worried as I don't understand why there are two arrays, and what exactly happened.
I know RAID 5 is problematic with such large disks, but that is nothing I can change at this time