[Gluster-users] slow seek times

[Osborne, Paul (paul.osborne@canterbury.ac.uk)]

Hi,

I have two identically specced web servers in VMware cloned from the same base image so all the packages etc are the same.

These are configured to mount the same volume from our Gluster Cluster (3.5.3) - which is fine.

What is interesting though is that the FS performance differs between the two.  I have run bonnie on both servers running the same tests as follows:

# bonnie -u root:root -x 5 -d /mnt/it-wiki -s 1024M:256 -r 512M

Whilst most of the bonnie output is near enough the same, the seeks and seeks_cpu numbers are noticeably different:

Server 1:   seeks: 3000.8                seeks_cpu: 9.6
Server 2:   seeks: 2175.4                seeks_cpu: 65.4

At this point I am confused why there should be such a difference in seek info and am uncertain how to proceed futher.

Suggestions are welcome.

Thanks

Paul

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20150513/a091e90f/attachment.html>