
Consolidating multiple SQL Server systems onto
Dell PowerEdge R610 Servers using Microsoft’s Hyper-V
21
User database files
These data files contain the data and objects for each user-defined
database. Their contents include database objects that you can
define, such as tables, indexes, and stored procedures.
When possible, group files with similar I/O characteristics. For
example, consolidate all logs. Because heterogeneous workloads
can have very different and even competing I/O characteristics,
combining heterogeneous workloads can hurt overall performance.
• BEST PRACTICE: When possible, separate the tempdb,
transaction log files, and user database files onto separate
disk groups.
Below, we illustrate the exact drive layout we used (see Figure 9).
We connected the Ethernet 0 port of the primary control module
for each EqualLogic PS5000 to the gigabit switch. We also
connected the gigabit switch to one of the Broadcom NICs in the
PowerEdge R610, which was dedicated to iSCSI traffic.
For detailed instructions on how we configured the storage, see the
section Creating the RAID group for the host OS in Appendix B
, as
well as Appendix C
.
Figure 9. The physical drive configuration we used in the
virtualization consolidated server in our hands-on testing and
research for this Guide.
Internal disk drives
Our PowerEdge R610 server contained six drives. We configured
these as a single RAID 10 for the parent partition, which holds the
host OS. We stored the VMs on the EqualLogic SAN.
Physical Disk Configuration - PowerEdge R610 with
EqualLogic PS5000
RAID 10 (Hyper-V
parent partition)
RAID 10 (VHD files for VMs)
Each VM had separate VHDs for:
OS and SQL Server software
DS2 database
Logs
Tempdb
PowerEdge R610
2 EqualLogic PS5000 iSCSI arrays
Comentarios a estos manuales