WAS: To use SAME or NOT for High End Storage Setup ?
- From: "Kevin Closson" <kevinc@xxxxxxxxxxxxx>
- To: <oracle-l@xxxxxxxxxxxxx>
- Date: Thu, 11 May 2006 13:09:56 -0700
There have been offline questions to me about the possible troubles with using a volume manager to stripe underlying RAID1+0 LUNS. This following is a link to the summary paper from a large test we did with IBM. The full paper is in the Collaborate proceedings and also available on our website (www.polyserve.com) under literature. This paper shows 80 spindles serving ~32,000 random 4KB xfers with 25% write to 8 nodes and then the array topped out (description of why in the paper). The file was in the CFS on a CVM volume consisting of 4 underlying RAID1+0 LUNS. The LUNS striped at 256KB (largest array supported) and the volume manager striped at 256KB to start and was switched to 1MB later for further testing. There was really no difference. http://www.polyserve.com/pdf/VMDB_BC_Summary_WP.pdf -- http://www.freelists.org/webpage/oracle-l
Other related posts:
- » WAS: To use SAME or NOT for High End Storage Setup ?