A Principled Technologies test report 9
Consolidating older database servers with the Intel processor-powered
Dell PowerEdge R730xd using SanDisk DAS Cache
APPENDIX B – HOW WE TESTED
About Benchmark Factory 7.0
Benchmark Factory® for Databases is a database performance testing tool that allows you to conduct database
workload replay, industry-standard benchmark testing, and scalability testing. Using the incorporated load testing tools,
you can make changes to your database environment, while mitigating the risks of unavoidable database changes such
as patches and upgrades, operating system migrations, and adjustments to virtual machine configurations. In addition,
its proprietary workload capture and replay supports Oracle and SQL Server database. For this study, we used the TPC-C
workload.
Learn more at software.dell.com/products/benchmark-factory/
About our testing
We used Benchmark Factory to create a 1,000-warehouse database, approximately 130 GB in size. We ran the
test with 5,000 users per database and default think times for 2 hours to ensure memory consumption in each
configuration. We used a 2,000-millisecond average response time as reported by Benchmark Factory as an indicator of
acceptable response and analyzed disk latencies to assess performance.
We configured the legacy IBM server with 48 GB of memory, and installed eight 300GB 10k disks for storage, the
default-shipping configuration of the server at its release. We configured two of those disks for a RAID 1 volume and set
up the remaining six disks with RAID 5 configuration. On the RAID 1 volume, we installed Windows Server 2003 R2 x64
and SQL Server 2008 R2, and stored SQL Server transaction logs. We used the RAID 5 volume to store the database files.
We configured the R730xd with 128 GB of DDR4 memory. For storage, we created a RAID 1 volume on the two
rear-facing HDDs to hold the operating system and database software, Windows Server 2012 R2 and SQL Server 2014.
We then configured six front HDDs, created a RAID 10 volume with no-read-ahead to store the database transaction
logs, and created a RAID 5 volume using 16 front HDDs to store the database files. For the SATA SSD DAS Cache
configuration, we enabled FastPath by creating a RAID 1 volume with the no-read-ahead and write-through enabled
policies, and used the two 800GB SATA SSD disks for that volume. We formatted only 400 GB of usable space to match
our PCIe configuration. For our PCIe DAS Cache configuration, we used two 400GB PCIe drives, mirrored using Windows
software RAID.
Running Benchmark Factory 7.0
We created a 1,000-warehouse TPC-C database on SQL Server 2008 R2, and then created a .bak file of the
database. We added this index before creating the backup file:
CREATE INDEX orders_nc1 ON c_order
(
o_w_id,
o_d_id,
o_c_id,
o_id
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, SORT_IN_TEMPDB = OFF, IGNORE_DUP_KEY = OFF,
DROP_EXISTING = OFF, ONLINE = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON MSSQL_fg
GO
Using that backup file and SQL Server’s built-in restore options, we were able to create the databases on each
server under test. We pointed Benchmark Factory clients hosted on load generation server, to each database on the
(45 Seiten)
Manymanuals.com
Manymanuals.de
Manymanuals.fr
Manymanuals.it
Manymanuals.pl
Manymanuals.cz
Manymanuals.es
Manymanuals-pt.com
Kommentare zu diesen Handbüchern