> your i/o must be really random to be seeing numbers that lousy (10
> seconds to replay a file is 1.6 megabytes/sec), or there is some other
> unexplained problem with your server. is your raid controller
> properly caching wites? have you benchmarked the volume with bonnie++
> or similar tool (pay close attention to seeks).
>
Here's the bonnie++ output (two runs):
Version 1.03 ------Sequential Output------ --Sequential Input-
--Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block--
--Seeks--
Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec
%CP
dev-db-2 32G 43174 99 87421 24 45614 12 48302 97 164574 23 205.3
0
------Sequential Create------ --------Random
Create--------
-Create-- --Read--- -Delete-- -Create-- --Read---
-Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec
%CP
16 +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++
+++
Having never used bonnie++ before, I don't have a baseline to compare this
against, but that looks like 87MB/s writes and 164MB/s reads to me. Am I
reading this correctly? It looks pretty good to me.
Here is some output from iostat
Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn
sda 0.00 0.00 0.00 0 0
sdb 1.00 0.00 55.72 0 112
sdc 1.00 0.00 63.68 0 128
sdd 101.49 1699.50 0.00 3416 0
avg-cpu: %user %nice %system %iowait %steal %idle
0.06 0.00 0.06 12.37 0.00 87.51
Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn
sda 2.50 0.00 24.00 0 48
sdb 0.00 0.00 0.00 0 0
sdc 42.50 0.00 8288.00 0 16576
sdd 101.50 1688.00 0.00 3376 0
avg-cpu: %user %nice %system %iowait %steal %idle
0.12 0.00 0.06 12.35 0.00 87.46
Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn
sda 0.00 0.00 0.00 0 0
sdb 0.00 0.00 0.00 0 0
sdc 0.00 0.00 0.00 0 0
sdd 112.44 1787.06 0.00 3592 0
avg-cpu: %user %nice %system %iowait %steal %idle
0.12 0.00 0.06 12.36 0.00 87.45
Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn
sda 0.00 0.00 0.00 0 0
sdb 4.50 0.00 48.00 0 96
sdc 0.50 0.00 4.00 0 8
sdd 97.50 1752.00 0.00 3504 0
In the above: sdb holds the pg_xlog directory, sdc holds the wal archive,
and sdd is the 4 disk RAID 1+0 where the pgdata directory is stored. All
these disks are ext3 with noatime,data=writeback mount options. The RAID
controller is an Adaptec 3805 with 128MB battery backed cache (only option
offered by our hosting provider for this server class).
Does any of this shed any light on how to boost my restore performance?
thanks,
Mason
