MirrorUX disk problem

[saqi]

Hi All,

We observed the Mirrored disks usage for a few days and found something strange. It seems that most of the time the mirror disk c4t10d0 at times is ~90% busy and relatively busier than the c4t8d0 disk.Below is an example .Please check whether there is a problem with the mirror disk or its normal behavior of disks.

# sar -d 4 5

HP-UX test B.11.11 U 9000/800 09/18/07

15:25:05 device %busy avque r+w/s blks/s avwait avserv
15:25:10 c2t1d0 25.70 0.50 43 478 0.00 8.76
c2t0d0 16.93 0.50 30 411 0.00 7.42
c4t8d0 45.02 0.50 124 4688 0.00 5.72
c4t10d0 94.82 0.50 165 4839 0.00 9.76
c4t12d0 1.39 0.50 2 37 0.00 5.47
15:25:15 c2t1d0 28.60 0.50 44 484 0.00 8.76
c2t0d0 19.40 0.50 32 428 0.00 7.61
c4t8d0 53.00 0.50 140 5270 0.00 6.13
c4t10d0 94.80 0.50 180 5417 0.00 9.26
c4t12d0 1.20 0.50 2 35 0.00 5.89
15:25:20 c2t1d0 27.15 0.50 45 513 0.00 9.21
c2t0d0 17.76 0.50 31 450 0.00 7.72
c4t8d0 55.09 0.50 145 6304 0.00 6.17
c4t10d0 93.61 0.50 177 6433 0.00 9.00
c4t12d0 1.80 0.50 3 45 0.00 6.54
15:25:25 c2t1d0 28.00 0.50 40 407 0.00 9.49
c2t0d0 17.40 0.50 28 349 0.00 7.13
c4t8d0 44.80 0.50 124 5003 0.00 5.54
c4t10d0 94.80 0.50 166 5171 0.00 9.31
c4t12d0 2.20 0.50 2 29 0.00 21.44
15:25:30 c2t1d0 23.85 0.52 36 351 0.12 10.53
c2t0d0 14.23 0.52 26 307 0.29 8.87
c4t8d0 50.30 0.50 141 5897 0.00 5.79
c4t10d0 94.59 0.50 177 6041 0.00 8.93
c4t12d0 0.80 0.50 2 38 0.00 3.89

Average c2t1d0 26.66 0.50 42 447 0.02 9.31
Average c2t0d0 17.15 0.50 29 389 0.05 7.72
Average c4t8d0 49.64 0.50 135 5432 0.00 5.88
Average c4t10d0 94.52 0.50 173 5580 0.00 9.25
Average c4t12d0 1.48 0.50 2 37 0.00 7.96

# strings /etc/lvmtab
/dev/vg00
/dev/dsk/c2t1d0
/dev/dsk/c2t0d0
/dev/vg01
/dev/dsk/c4t8d0
/dev/dsk/c4t10d0
/dev/vg02
/dev/dsk/c4t12d0

Thanks

[Steven E. Protter]

Shalom,

You may have a problem with the mirror disk that is at 90%

dmesg

cstm or mstm or xstm to run diagnosis.

vgdisplay -v vg00

A disk is bad or in trouble.

http://docs.hp.com/en/5991-1236/When_Good_Disks_Go_Bad.pdf

SEP

[Duncan Edmonstone]

It's nothing to worry about - reads won't show against both disks so you usually have one disk busier than the other. Note a similar situation for the disks in vg00.

HTH

Duncan