SAN Layout Question...

Greetings, I am working with IFX 10UC4 and migrating data to an EMC SAN. We currently have 2 LUNS available for use with our IFX database. The LUNS are setup to use RAID 1+0. Machine has 8 dual core processors. The database is approximately 300GB, however, through the use of HPL and some SQL we have reduced the amount of data to < 200GB. The disk layout proposed is as follows: LUN1 - 260GB dbs10 dbs08 dbs06 dbs04 dbs02 tempsp02 tempsp04 TXLogs rootdbs LUN2 - 440GB dbs09 dbs07 dbs05 dbs03 dbs01 tempsp01 tempsp03 TXLogs When using HPL to laod the data into IFX on the SAN I am seeing an almost 100% utilization of LUN2 and only sporadic usage of LUN2 (iostat -x). The load job is taking an excessive amount of time, and, I was wondering if you had any advice regarding the layout detailed above. The CPUS's are loafing during the load but IOWAIT is ranging from 60 -90%. It seems as though the HPL jobs are writing 10K records from one file, writing 10K records from the second etc as opposed to running in parallel. The structures we are loading data to have been sized properly so as to have a minimum amount of extents (loaded the data first with no extent specified and looked at the space used etc...). I was thinking a better layout would be something along the lines of: LUN1 - 260GB tempsp01 tempsp02 tempsp03 tempsp04 TXLogs LUN2 - 440GB dbs10 dbs09 dbs08 dbs07 dbs06 dbs05 dbs04 dbs03 dbs02 dbs01 rootdbs Hoping someone can shed some light as to what an appropriate disk/space layout would be for this scenario?

OS is Redhat AS 4 "Tam O'Shanter" <tam@oshanter.com> wrote in message news:uTAhh.32357$6F1.8725@fe11.news.easynews.com... > Greetings, > > I am working with IFX 10UC4 and migrating data to an EMC SAN. > > We currently have 2 LUNS available for use with our IFX database. > > The LUNS are setup to use RAID 1+0. > > Machine has 8 dual core processors. > > The database is approximately 300GB, however, through the use of HPL and > some SQL we have reduced the amount of data to < 200GB. > > The disk layout proposed is as follows: > > LUN1 - 260GB > dbs10 > dbs08 > dbs06 > dbs04 > dbs02 > tempsp02 > tempsp04 > TXLogs > rootdbs > > LUN2 - 440GB > dbs09 > dbs07 > dbs05 > dbs03 > dbs01 > tempsp01 > tempsp03 > TXLogs > > When using HPL to laod the data into IFX on the SAN I am seeing an almost > 100% utilization of LUN2 and only sporadic usage of LUN2 (iostat -x). > > The load job is taking an excessive amount of time, and, I was wondering > if you had any advice regarding the layout detailed above. > > The CPUS's are loafing during the load but IOWAIT is ranging from 60 -90%. > > It seems as though the HPL jobs are writing 10K records from one file, > writing 10K records from the second etc as opposed to running in parallel. > > The structures we are loading data to have been sized properly so as to > have a minimum amount of extents (loaded the data first with no extent > specified and looked at the space used etc...). > > I was thinking a better layout would be something along the lines of: > > LUN1 - 260GB > tempsp01 > tempsp02 > tempsp03 > tempsp04 > TXLogs > > LUN2 - 440GB > dbs10 > dbs09 > dbs08 > dbs07 > dbs06 > dbs05 > dbs04 > dbs03 > dbs02 > dbs01 > rootdbs > > Hoping someone can shed some light as to what an appropriate disk/space > layout would be for this scenario? > > >