Linux Raid issues
 

I'm trying to build a simple Raid 5 array out of 4 SATA drives. They dries are , 250GB, 250GB, 500GB and 500GB. I've created partitions on each of them that are the exact same size, roughly 250GB.

/dev/sda1, /dev/sdb1. /dev/sdc1, /dev/sdd1

I created an array on these partiions before, but after a power outage it just died. not sure why...but since there was little of value on there, I just decided to rebuild the set from scratch. I first tried...

mdadm --create /dev/md0 --level=5 --raid-devices=4 /dev/sda1 /dev/sdb1 /dev/sdc1 /dev/sdd1

The device would start and then begin rebuilding as normal...about 5 minutes later it would show up like this.

+++++++++++++++++++++++++++++++++++++++
Version : 00.90.03
Creation Time : Wed Jan 16 21:03:10 2008
Raid Level : raid5
Array Size : 732563712 (698.63 GiB 750.15 GB)
Used Dev Size : 244187904 (232.88 GiB 250.05 GB)
Raid Devices : 4
Total Devices : 4
Preferred Minor : 0
Persistence : Superblock is persistent

Update Time : Wed Jan 16 21:05:22 2008
State : clean, degraded
Active Devices : 2
Working Devices : 3
Failed Devices : 1
Spare Devices : 1

Layout : left-symmetric
Chunk Size : 64K

UUID : 26ab635c:6dcc91db:645e3c10:269b8dbf
Events : 0.34

Number Major Minor RaidDevice State
0 8 33 0 active sync /dev/sdc1
1 8 17 1 active sync /dev/sdb1
2 0 0 2 removed
3 0 0 3 removed

4 8 49 - spare /dev/sdd1
5 8 1 - faulty spare /dev/sda1

+++++++++++++++++++++++++++++++++++++
So it looks like 2 drives are faulty...but I can't see that they are. I have tested both of the drives and they appear fine. I can mount them and use them as individual drives, but in the raid set it's not working.

I tried zeroing the superblocks on each and I've repartitioned each of the drives. still the same result.

any help is appreciated.

thorough drive testing?
 

When you indicate testing the drives, do you mean a SMART long test with no failure or just the measures mentioned in your post (mounting and building a filesystem)? These latter may not exercise the drive sufficiently to find an error. I've read some complaints of the Linux raid manager too easily marking a drive as faulty, but your circumstances (power outage, repeated recreate problems) suggest the likelihood of a true hardware issue. Check out the smartmontools package if these tests are not already in play.

If drive tests aren't an issue, you might check with a Linux RAID guru, perhaps via the appropriate mailing list, about the output of your mdadm. I'm not an expert, but with 4 physical devices and the requested RAID array of 4 devices, I don't see how there should ever be a spare (sdd1 in this case) unless this output is an artifact of being in the process of creating/rebuilding on top of devices of which some have pre-existing superblock info (the apparently good ones.)

I'll try that
 

I ran Bonnie++ just as a cursory test, but I will do some more digging per your suggestion.