Hi,
I used sysbasher and the web interface on the san. I figure
that it gives an ok view of the load on the san. It corresponds ok with timed
cp's.
Also i did some loops that just copied to and from the
san. But generally it's just a rough average, and i'm just wondering if
it's similar in performance to other configurations.
The hardware is:
2 fujitsusiemens RX200s3 servers with emulex 11003
hbas and two dualcore cpus and 4gb memory.
Connected is a SX60 san with an extra controller and
powersupply.
Mvh / Kind regards
Kristoffer Lippert
Systemansvarlig
JP/Politiken A/S
Online
Magasiner
Tlf. +45 8738 3032
Cell. +45 6062 8703
Fra: linux-cluster-bounces@xxxxxxxxxx [mailto:linux-cluster-bounces@xxxxxxxxxx] På vegne af Dan.Askew@xxxxxxxxxxxxx
Sendt: 31. juli 2007 12:12
Til: linux clustering
Emne: Re: SV: Linux Clustering Newbe
No I do not but if you elaborate on the tests performed or have atest script I would be glad run some bench marks.
It would also be helpful to know the hardware you are running on.
Dan Askew
Sr Systems Administrator
The J. M. Smucker Company
Phone (330)684-3662
"Kristoffer Lippert"
<kristoffer.lippert@xxxxxxxx> Sent by: linux-cluster-bounces@xxxxxxxxxx 07/31/2007 04:32 AM
|
|
Hi Dan,
Have you got any performance figures on the OCFS2 ?
My GFS is currently delivering the following performance:
(2 nodes, 2gbit FC HBAs multipath'ed + 1 Fujitsu SAN (SX60) 7x250gb raid 5.)
Writes: average 100mb/sec, 200 files/sec (with simple cp/rm)
Writes: max 150mb/sec, 250files/sec (with simple cp/rm)
Reads: average 160mb/sec (cat > dev/null - uncached)
Reads: average 1150mb/sec (cat > dev/null - cached)
Anybody have a feeling about this performance level?
Regards
Kristoffer
Fra: linux-cluster-bounces@xxxxxxxxxx [mailto:linux-cluster-bounces@xxxxxxxxxx] På vegne af Dan.Askew@xxxxxxxxxxxxx
Sendt: 20. juli 2007 16:19
Til: linux clustering
Emne: Linux Clustering Newbe
Ok Here is what I did to get it working:
used RHCS to manage the cluster services
used CLVM to manage the LVM metadata across cluster
Coded a "locking_type = 3" in the LVM.CONF file
Used OCFS2 to cluster and manage the filesystems across cluster.
Configuration is
2 node cluster runing nfs
vitural IPAddr service to allow clients to mount the drives via NFS
All appears to be working fine. I can fail over the services manually or even boot a server and the failover works great.
Now the big question....is this supported by redhat and oracle ...hmmmm
Thanks to all who helped get this working...
--
Linux-cluster mailing list
Linux-cluster@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/linux-cluster
-- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster