I had a bad sector on a raid6 member disk, fixed it with dd and then
decided to bite the bullet and upgrade to 2.6.16rc4 to get the better
bad sector error handling.
I note this kernel has the "fix raid6 resync check/repair code" patch.
Since my disk was out of the array briefly I expected to see
reduced writes during resync but I am not seeing this. In fact,
the syncing disk doesnt get read at all!
Here's an iostat output:
Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn
sdd 43.58 8989.34 4.25 12172822 5752
sdh 43.01 8988.78 3.17 12172070 4288
sdi 46.17 8982.13 1.01 12163062 1368
sdj 48.42 8989.03 3.23 12172408 4376
sdk 46.39 2.56 8963.30 3470 12137568
sdo 50.18 8986.71 2.98 12169268 4032
sds 50.71 8992.34 3.85 12176890 5218
sdt 48.61 8989.37 4.22 12172868 5720
(I've deleted the uninteresting disks)
Is it a pilot error? I added the missing disk (sdk) to the array
and mdadm said 're-adding' IIRC. I have no bitmaps yet in case that matters.
/dev/md1:
Version : 00.90.03
Creation Time : Mon Dec 5 08:38:36 2005
Raid Level : raid6
Array Size : 1757815296 (1676.38 GiB 1800.00 GB)
Device Size : 292969216 (279.40 GiB 300.00 GB)
Raid Devices : 8
Total Devices : 8
Preferred Minor : 1
Persistence : Superblock is persistent
Update Time : Fri Feb 24 09:12:32 2006
State : clean, degraded, recovering
Active Devices : 7
Working Devices : 8
Failed Devices : 0
Spare Devices : 1
Chunk Size : 128K
Rebuild Status : 2% complete
UUID : 628e4bb1:71f6dbea:5b4cd026:7ac5282f
Events : 0.2285138
Number Major Minor RaidDevice State
0 8 49 0 active sync /dev/sdd1
1 65 49 1 active sync /dev/sdt1
2 65 35 2 active sync /dev/sds3
8 8 161 3 spare rebuilding /dev/sdk1
4 8 225 4 active sync /dev/sdo1
5 8 113 5 active sync /dev/sdh1
6 8 145 6 active sync /dev/sdj1
7 9 4 7 active sync /dev/md4
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
