[Gluster-devel] Performance

[Anand Avati]

Brian,
 block size of 1kb is too small and expensive, especially for network or
fuse based filesystems. Please try with a larger block size like 64kb.

avati

On 10/14/07, Brian Taber <btaber at diversecg.com> wrote:
>
> I am new to ClusterFS and I am looking for a replacement for my NFS setup.
> I have configured a clusterfs server on top of a raid 5 array on 4 SATA
> hard drives.  Directly I can get a speed of 72.3 MB/s weh I do a:
>
> dd if=/dev/zero bs=1024 count=1000000 of=/data/1Gb.file
>
> If I do the same test over NFSv3 I get performance of 14.1 MB/s
>
> If I do the same test over the gluster mount, I get perfromance of 3.6MB/s
>
> Am I doing something wrong here?  How can I increase my performance to the
> same or beyond my current NFS?
>
> I setup the server with this config:
>
> volume brick-ns
>   type storage/posix
>   option directory /gluster-ns
> end-volume
>
> volume brick
>   type storage/posix
>   option directory /data/gluster
> end-volume
>
> volume iothreads1    #iothreads can give performance a boost
>    type performance/io-threads
>    option thread-count 8
>    subvolumes brick
> end-volume
>
> volume server
>   type protocol/server
>   subvolumes iothreads1 brick-ns
>   option transport-type tcp/server     # For TCP/IP transport
>   option auth.ip.iothreads1.allow 192.168.*
>   option auth.ip.brick-ns.allow 192.168.*
> end-volume
>
>
> and setup a client with:
>
> volume client1-ns
>   type protocol/client
>   option transport-type tcp/client
>   option remote-host 192.168.200.201
>   option remote-subvolume brick-ns
> end-volume
>
> volume client1
>   type protocol/client
>   option transport-type tcp/client
>   option remote-host 192.168.200.201
>   option remote-subvolume iothreads1
> end-volume
>
> volume bricks
>   type cluster/unify
>   subvolumes client1
>   option namespace client1-ns
>   option scheduler alu
>   option alu.limits.min-free-disk  60GB              # Stop creating files
> when free-space lt 60GB
>   option alu.limits.max-open-files 10000
>   option alu.order
> disk-usage:read-usage:write-usage:open-files-usage:disk-speed-usage
>   option alu.disk-usage.entry-threshold 2GB          # Units in KB, MB and
> GB are allowed
>   option alu.disk-usage.exit-threshold  60MB         # Units in KB, MB and
> GB are allowed
>   option alu.open-files-usage.entry-threshold 1024
>   option alu.open-files-usage.exit-threshold 32
>   option alu.stat-refresh.interval 10sec
> end-volume
>
> volume writebehind   #writebehind improves write performance a lot
>   type performance/write-behind
>   option aggregate-size 131072 # in bytes
>   subvolumes bricks
> end-volume
>
>
> When trying the
>
>
>
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at nongnu.org
> http://lists.nongnu.org/mailman/listinfo/gluster-devel
>



-- 
It always takes longer than you expect, even when you take into account
Hofstadter's Law.

-- Hofstadter's Law