[Gluster-devel] Performance

Roland Fischer roland.fischer at xidras.com
Thu Mar 18 07:16:08 UTC 2010


hi Raghavendra G,

i use xen 3.4.1 on glusterfs. the images of virtual maschine are on 
glusterfs hosted. i run 16 domUs (xen images) on glusterfs an the disk 
speed is awful slow. and in monitoring tool i see that there is a lot of 
wait cpu (that means that the cpus of the domUs are waiting on 
glusterfs). before i switched on glusterfs the domU images laying on 
local disk and there are no wait cpu....

best regards roland

Am 18.03.2010 05:46, schrieb Raghavendra G:
> Hi Roland,
>
> What are the applications you are using on glusterfs? In particular, 
> what is the i/o pattern of applications? As a general guideline, you 
> can try enabling/disabling each of the performance translators and 
> observe the gain/loss of performance and tune the configuration 
> accordingly.
>
> regards,
> On Wed, Mar 17, 2010 at 9:05 PM, Roland Fischer 
> <roland.fischer at xidras.com <mailto:roland.fischer at xidras.com>> wrote:
>
>     Hi Community,
>
>     i need your help. i have performance problems with glusterfs 3.0.0
>     and domUs (xen)
>
>     i use 2 ident glusterfs-server (physikal HW) and two xen-server
>     (physikal)
>
>     currently i use client side replication - which is awful slow. i
>     use a monitoring tool and see in domUs that there is a lot of cpu
>     waiting. (before i switch to glusterfs there was no wait CPU)
>
>     is server-side-replication faster and failsave. i mean if one
>     glusterfs server goes down, does the other take over the domUs?
>
>     is there anything in volfiles which i can tune?!? should i use
>     server-side-replication?!?
>
>     should i use the --disable-direct-io-mode? if yes on server side
>     or client-side or both  -  and how to add in fstab (with
>     --disable-direct-io-mode)?????
>
>     Thank you for your help!!!
>
>     servervolfile:
>     cat /etc/glusterfs/export-domU-images-client_repl.vol
>     #############
>     volume posix
>      type storage/posix
>      option directory /GFS/domU-images
>     end-volume
>
>     volume locks
>      type features/locks
>      subvolumes posix
>     end-volume
>
>     volume domU-images
>      type performance/io-threads
>      option thread-count 8 # default is 16
>      subvolumes locks
>     end-volume
>
>     volume server
>      type protocol/server
>      option transport-type tcp
>      option auth.addr.domU-images.allow 192.*.*.*,127.0.0.1
>      option transport.socket.listen-port 6997
>      subvolumes domU-images
>     end-volume
>     ######################
>
>     clientvolfiles:
>
>     cat /etc/glusterfs/mount-domU-images-client_repl.vol
>     volume gfs-01-01
>      type protocol/client
>      option transport-type tcp
>      option remote-host hostname
>      option transport.socket.nodelay on
>      option remote-port 6997
>      option remote-subvolume domU-images
>      option ping-timeout 5
>     end-volume
>
>     volume gfs-01-02
>      type protocol/client
>      option transport-type tcp
>      option remote-host hostname
>      option transport.socket.nodelay on
>      option remote-port 6997
>      option remote-subvolume domU-images
>      option ping-timeout 5
>     end-volume
>
>     volume gfs-replicate
>      type cluster/replicate
>      subvolumes gfs-01-01 gfs-01-02
>     end-volume
>
>     volume writebehind
>      type performance/write-behind
>      option cache-size 4MB   #default 16
>      subvolumes gfs-replicate
>     end-volume
>
>     volume readahead
>      type performance/read-ahead
>      option page-count 8              # cache per file  = (page-count
>     x page-size)
>      subvolumes writebehind
>     end-volume
>
>     volume iocache
>      type performance/io-cache
>      option cache-size 1GB   #new 1GB supported
>      option cache-timeout 1
>      subvolumes readahead
>     end-volume
>
>     volume statprefetch
>        type performance/stat-prefetch
>        subvolumes iocache
>     end-volume
>
>     #################################################
>
>
>     Best regards,
>     Roland
>
>
>
>     _______________________________________________
>     Gluster-devel mailing list
>     Gluster-devel at nongnu.org <mailto:Gluster-devel at nongnu.org>
>     http://lists.nongnu.org/mailman/listinfo/gluster-devel
>
>
>
>
> -- 
> Raghavendra G
>






More information about the Gluster-devel mailing list