![]() | |
![]() |
| | Thread Tools | Display Modes |
#21
| |||
| |||
|
|
I'm attempting to set up a 2 node active/active sql 2005 (ent edition) failover cluster with an HP 2000i MSA iscsi storage box. Server is 2003 Ent Edition. At the hardware level on the 2000i, I've created 3 raid 10 virtual disks as follows: SqlData w/ 2 volumes, SqlData1 and SqlData2 SqlLog w/ 2 volumes, SqlLog1 and SqlLog2 SqlTempDB, w/ 3 volumes, Quorum, TempDB_DB01 and TempDB_02. The intent is for one active node to 'own' the SqlData1, SqlLog1, and TempDB_DB01 volumes and the other node would own SqlData2, SqlLog2, and Tempdb_02 - with the quorum volumne serving as the quorum. It looks like at this same hardware level I can map these volumes to the physical db machines via luns. My first question is: Do I use the same lun number for mapping a volume to the physical host? i.e, when mapping SQlData1 to physical host sqldb01 I am using lun # 6 - when I map the same volume to physical host sqldb02 do I use lun # 6 again or does it need to be a different lun? I assume BOTH physical db nodes need to be mapped for failover to work. Second question, similar to the first - on each host do the drive letters for these volumes need to be the same or should they be different? i.e., continuing from teh example above I have the SqlData1 volume on physical host sqldb01 using drive letter 'Y' - would I use the *same* drive letter on physical host sqldb02 or should it be different? Thanks in advance, I'm running out of time to get this going ![]() |
![]() |
| Thread Tools | |
| Display Modes | |
| |