I have recently taken over administration of a MySQL database server and from some performance monitoring I have noticed heavy disk usage.


Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s avgrq-sz avgqu-sz await svctm %util
sda 1.27 64.50 7.90 71.55 634.65 1112.11 21.99 0.06 2.45 2.01 15.99
sdb 0.31 57.29 1.25 2.78 190.25 480.57 166.12 0.06 15.08 2.40 0.97
sdc 0.00 0.00 0.00 0.00 0.01 0.00 27.08 0.00 1.64 1.62 0.00
sdd 0.01 87.38 14.54 5.15 119.11 740.26 43.64 0.01 0.59 0.40 0.78

Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s avgrq-sz avgqu-sz await svctm %util
sda 0.00 119.60 1.40 265.00 11.20 3155.20 11.89 49.87 184.87 2.09 55.60
sdb 0.00 100.60 0.00 2.40 0.00 824.00 343.33 0.03 11.50 3.17 0.76
sdc 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
sdd 0.00 140.00 79.40 7.20 635.20 1177.60 20.93 0.03 0.34 0.26 2.26

Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s avgrq-sz avgqu-sz await svctm %util
sda 0.00 114.77 1.40 270.86 11.18 3113.77 11.48 61.05 220.76 1.99 54.15
sdb 0.00 479.64 0.80 10.18 6.39 3918.56 357.53 0.17 15.89 3.40 3.73
sdc 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
sdd 0.00 281.44 110.58 6.99 884.63 2307.39 27.15 0.07 0.61 0.43 5.07

The current configuration is:

3x 76Gb + 3x 146GB 10K RPM SAS drives
* 2x 146GB Raid 1 for SDA (OS) and MySQL data
* 1x 76GB Non-Raid SDB (InnoDB table space)
* 1x 76GB Cold Standby (SDC)
* 1x 76GB Non-Raid SDD (Binary logs)
* 1x 146GB Hot-Standby

Given the disk IO I have suggested that we re-configure to the following:

2x 76GB in RAID 1 for OS and Binary Logs
4x 146GB in RAID 10 for InnoDB and MySQL Data

In an effort to strengthen my case, how much improvement (if any) could I expect from the proposed configuration change? I am already aware of the obvious redundancy and fail over from the new configuration so just looking at this from a performance perspective.

Any other suggestions on improving disk IO is also welcomed.

Thanks

Gavin