[Gluster-users] Is that iozone result normal?

[Kirby Zhou]

5-nodes server and 1 node client are connected by gigabits Ethernet.

#] iozone -r 32k -r 512k -s 8G
              KB  reclen   write rewrite    read    reread    read   write
read rewrite    read   fwrite frewrite   fread  freread
         8388608      32   10559    9792    62435    62260
         8388608     512   63012   63409    63409    63138

It seems 32k write/rewrite performance are very poor, which is different
from a local file system.
In the local file system, the big difference will only present on the random
read/write stage.

========================
Server conf:
========================
volume brick1-raw
  type storage/posix                   # POSIX FS translator
  option directory /exports/disk1      # Export this directory
end-volume

volume brick1
  type performance/io-threads
  subvolumes brick1-raw
  option thread-count 16
  option cache-size 256m
end-volume

volume brick2-raw
  type storage/posix                   # POSIX FS translator
  option directory /exports/disk2      # Export this directory
end-volume

volume brick2
  type performance/io-threads
  subvolumes brick2-raw
  option thread-count 16
  option cache-size 256m
end-volume

volume brick-ns
  type storage/posix                   # POSIX FS translator
  option directory /exports/ns         # Export this directory
end-volume========================
Client conf:
========================
volume unify0-raw
  type cluster/unify
  subvolumes remote1-brick1 remote1-brick2 remote2-brick1 remote2-brick2
remote3-brick1 remote3-brick2 remote4-brick1 remote4-brick2 remote5-brick1
remote5-brick2
  option namespace afr-ns0
  option scheduler alu
  option alu.limits.min-free-disk  5%
  option alu.order disk-usage:open-files-usage
  option alu.disk-usage.entry-threshold 1GB   # Kick in if the discrepancy
in disk-usage between volumes is more than 1GB
  option alu.disk-usage.exit-threshold  60MB   # Don't stop writing to the
least-used volume until the discrepancy is 1988MB 
  option alu.open-files-usage.entry-threshold 1024   # Kick in if the
discrepancy in open files is 1024
  option alu.open-files-usage.exit-threshold 32   # Don't stop until 992
files have been written the least-used volume
end-volume

volume unify0-io-cache
  type performance/io-cache
  option cache-size 256MB            # default is 32MB
  option page-size 1MB               # 128KB is default option
  #option priority *.h:3,*.html:2,*:1 # default is '*:0'
  #option force-revalidate-timeout 2  # default is 1
  subvolumes unify0-raw
end-volume

volume unify0-writebehind
  type performance/write-behind
  option aggregate-size 1MB      # default is 0bytes
  #option window-size 3MB         # default is 0bytes
  option flush-behind on         # default is 'off'
  subvolumes unify0-io-cache
end-volume

volume unify0
  type features/fixed-id
  option fixed-uid 99 
  option fixed-gid 99
  subvolumes unify0-io-cache
end-volume