Disk performance - avqueue

[Mauro Livi]

Hi all,
We have a RP3440 server running 11.11 that is essentially our application server. It has internal disks but one is used for mirroring. We have the OS plus the apps programs(oracle financials and some customized COBOL programs) that we run on the same disk...I know, I know, these should be separated, but that's a discussion for another day.

Anyway, if I do a sar -d I do see avqueue numbers go above 1 periodically, however, avwait times are always below avserv times and the disks are usually not that busy (see below). Not much writing is going on here (like I said, this is an app server) so are the avqueue numbers just a byproduct of having the OS and the apps being on the same disk?
People are not squawking about performance or anything and CPU and memory on this server is always good (i.e. no bottlenecks there). I just wondered if there is something I can do about these avqueue numbers and if I can get them below 1.
Many thanks
Mauro

08:13:33 device %busy avque r+w/s blks/s avwait avserv
08:13:43 c2t0d0 5.00 0.50 10 89 0.00 11.80
c2t1d0 3.40 0.50 7 78 0.00 8.21
08:13:53 c2t0d0 4.10 0.50 8 70 0.00 10.50
c2t1d0 2.90 0.50 6 62 0.00 8.98
08:14:03 c2t0d0 2.20 0.50 4 25 0.00 7.05
c2t1d0 1.70 0.50 3 20 0.00 7.37
08:14:13 c2t0d0 3.40 0.50 7 61 0.00 8.95
c2t1d0 2.30 0.50 5 52 0.00 6.94
08:14:23 c2t0d0 1.90 0.50 3 23 0.00 7.03
c2t1d0 1.30 0.50 2 19 0.00 6.22
08:14:33 c2t0d0 9.40 2.92 25 199 6.01 20.88
c2t1d0 6.90 2.16 20 180 4.91 20.09
08:14:43 c2t0d0 1.70 0.50 3 27 0.00 7.03
c2t1d0 1.10 0.50 2 25 0.00 6.21
08:14:53 c2t0d0 1.50 0.50 2 23 0.00 7.68
c2t1d0 1.10 0.50 2 20 0.00 6.55
08:15:03 c2t0d0 3.20 0.50 8 40 0.00 5.17
c2t1d0 2.00 0.50 6 35 0.00 3.56
08:15:13 c2t0d0 3.10 0.50 6 47 0.00 7.25
c2t1d0 4.50 0.50 4 39 0.00 33.59
08:15:23 c2t0d0 1.60 0.50 3 18 0.00 8.33
c2t1d0 1.30 0.50 2 15 0.00 7.02
08:15:33 c2t0d0 2.70 0.50 5 39 0.00 8.43
c2t1d0 2.10 0.50 4 33 0.00 7.84
08:15:43 c2t0d0 5.51 0.50 17 348 0.00 5.38
c2t1d0 3.80 0.50 7 80 0.00 9.45
08:15:53 c2t0d0 4.10 0.50 9 71 0.00 10.68
c2t1d0 3.40 0.50 6 52 0.00 11.84
08:16:03 c2t0d0 4.10 0.50 10 69 0.00 9.35
c2t1d0 3.40 0.50 7 60 0.00 10.01
08:16:13 c2t0d0 3.90 0.50 8 85 0.00 6.41
c2t1d0 2.60 0.50 5 55 0.00 7.82
08:16:23 c2t0d0 3.30 0.50 6 37 0.00 7.97
c2t1d0 2.10 0.50 3 28 0.00 7.93
08:16:33 c2t0d0 5.99 1.24 15 110 2.21 17.60
c2t1d0 4.60 1.39 12 100 2.59 18.08
08:16:43 c2t0d0 5.91 0.50 11 105 0.00 11.20
c2t1d0 3.70 0.50 7 91 0.00 10.81
08:16:53 c2t0d0 5.29 0.53 11 75 0.21 15.26
c2t1d0 3.80 0.51 9 67 0.14 14.81
08:17:03 c2t0d0 3.80 0.50 6 44 0.00 9.53
c2t1d0 2.30 0.50 5 36 0.00 8.80
08:17:13 c2t0d0 3.00 0.50 5 47 0.00 8.63
c2t1d0 1.80 0.50 4 39 0.00 6.63
08:17:23 c2t0d0 1.80 0.50 3 17 0.00 7.91
c2t1d0 1.10 0.50 2 14 0.00 6.13
08:17:33 c2t0d0 2.80 0.50 5 30 0.00 7.99
c2t1d0 1.90 0.50 4 24 0.00 5.77
08:17:43 c2t0d0 4.50 0.50 8 66 0.00 8.73
c2t1d0 2.40 0.50 5 55 0.00 7.12
08:17:53 c2t0d0 3.30 0.50 6 37 0.00 9.81
c2t1d0 2.40 0.50 4 31 0.00 9.19
08:18:03 c2t0d0 4.20 0.50 10 153 0.00 7.39
c2t1d0 2.70 0.50 5 45 0.00 8.46
08:18:13 c2t0d0 4.00 0.50 7 65 0.00 8.28
c2t1d0 2.70 0.50 5 55 0.00 7.25
08:18:23 c2t0d0 4.00 0.50 8 126 0.00 8.02
c2t1d0 2.30 0.50 4 40 0.00 7.80
08:18:33 c2t0d0 4.30 0.50 9 57 0.00 8.74
c2t1d0 3.00 0.50 7 49 0.00 6.48
08:18:43 c2t0d0 5.21 0.50 11 98 0.00 11.26
c2t1d0 3.60 0.50 8 81 0.00 9.28
08:18:53 c2t0d0 5.29 0.50 12 91 0.00 15.34
c2t1d0 4.30 0.50 10 81 0.00 15.72
08:19:03 c2t0d0 3.20 0.50 7 46 0.00 9.14
c2t1d0 2.30 0.50 5 39 0.00 8.43
08:19:13 c2t0d0 3.50 0.50 7 68 0.00 9.10
c2t1d0 2.10 0.50 5 60 0.00 7.39
08:19:23 c2t0d0 3.60 0.50 11 231 0.00 4.71
c2t1d0 2.00 0.50 4 64 0.00 6.71
08:19:33 c2t0d0 3.30 0.50 6 79 0.00 7.80
c2t1d0 2.80 0.50 5 71 0.00 7.98
08:19:43 c2t0d0 4.30 0.50 9 120 0.00 8.13
c2t1d0 3.40 0.50 6 109 0.00 8.76
08:19:53 c2t0d0 3.30 0.50 6 75 0.00 9.08
c2t1d0 2.60 0.50 5 68 0.00 8.90
08:20:03 c2t0d0 4.70 0.50 10 115 0.00 6.81
c2t1d0 3.90 0.50 8 100 0.00 6.37
08:20:13 c2t0d0 3.70 0.50 7 95 0.00 7.14
c2t1d0 2.50 0.50 5 85 0.00 7.18
08:20:23 c2t0d0 2.20 0.50 4 71 0.00 8.10
c2t1d0 1.80 0.50 3 67 0.00 7.94
08:20:33 c2t0d0 2.40 0.50 5 47 0.00 7.00
c2t1d0 2.00 0.50 3 42 0.00 7.85
08:20:43 c2t0d0 6.80 0.50 9 79 0.00 13.88
c2t1d0 3.00 0.50 6 67 0.00 7.84
08:20:53 c2t0d0 4.10 0.50 7 52 0.00 10.10
c2t1d0 2.60 0.50 5 44 0.00 8.55
08:21:03 c2t0d0 3.00 0.50 6 41 0.00 10.65
c2t1d0 2.10 0.50 4 36 0.00 7.80
08:21:13 c2t0d0 3.40 0.50 6 47 0.00 9.07
c2t1d0 2.20 0.50 4 39 0.00 6.52
08:21:23 c2t0d0 3.00 0.50 8 133 0.00 5.28
c2t1d0 1.80 0.50 3 27 0.00 7.37
08:21:33 c2t0d0 3.60 0.50 7 57 0.00 10.04
c2t1d0 2.20 0.50 4 36 0.00 7.43
08:21:43 c2t0d0 5.00 0.50 10 153 0.00 9.46
c2t1d0 3.30 0.50 7 129 0.00 11.15
08:21:53 c2t0d0 3.80 0.50 7 50 0.00 11.02
c2t1d0 2.80 0.50 5 43 0.00 9.00

Average c2t0d0 3.82 0.68 8 79 0.47 10.11
Average c2t1d0 2.68 0.67 5 56 0.49 10.29

[Steven E. Protter · Accepted solution]

Shalom,

If there are no response complaints, the data should be recorded for future reference. The first rule of thumb to avoid frustration is not to react to numbers by themselves.

If there is a response issue, you may have to change the application layout and perhaps try and spread the i/o between disks.

But if there are no user complaints, I would not make a great deal of extra work for myself.

SEP

[Mauro Livi]

Thanks SEP...yeah, kind of what I figured. I don't particularly care for apps and the OS being on the same disk, but that wasn't my call.
Users aren't complaining so I've basically documented this and told my boss...kind of a CYA move :)
Thanks again
Mauro