[Gluster-users] performance lower then expected

[Michael Di Domenico]

No, write-behind is not enabled.

The volume spec files are pretty much exactly the same as the ones on the
wiki page, all I changed was the directory.  (I can't post them as i'm
travelling at the moment)



On Sat, Oct 25, 2008 at 12:49 AM, Anand Avati <avati at zresearch.com> wrote:

> Are you using write-behind? Can you post your volume spec files?
>
> avati
>
> 2008/10/24, Michael Di Domenico <mdidomenico4 at gmail.com>:
>>
>> I've setup an eight node server stripe using gluster 1.4.0pre5 using the
>> stripe example from the wiki.  Each of these eight nodes has a 100Mbit
>> ethernet card and a single hard disk.
>> I've connected them all together using a gigabit switch and I have a
>> gigabit workstation connected, with gluster mounted and running fine.
>>
>> However, when i try to do a dd test to the disk "dd if=/dev/zero of=test1
>> bs=1024k count=1000", I'm only seeing 10MB/sec over the transfer.
>>
>> I would suspect the data to stripe evenly across all the disks/nodes,
>> effectively giving me at least over 50MB/sec in combined throughput.
>>
>> I'm I confused on what the stripe translator does, or did I do something
>> wrong?
>>
>>
>> _______________________________________________
>> Gluster-users mailing list
>> Gluster-users at gluster.org
>> http://zresearch.com/cgi-bin/mailman/listinfo/gluster-users
>>
>>
>
>
> --
> If I traveled to the end of the rainbow
> As Dame Fortune did intend,
> Murphy would be there to tell me
> The pot's at the other end.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20081028/7881af2c/attachment.html>