White Papers

BP1021 Microsoft SQL Server 2008 Backup and Restore using Dell EqualLogic 22
4.2.2 Test #2: Performance impact studies during full, tlog, and differential backups
This test simulates an environment with weekly full backups, daily differential backups, and transaction
log backups for every 30 minutes assuming the typical production duration to be eight hour workday.
• To simulate this test, we ran the user load from Benchmark Factory for ten hours, scheduled a
o Full backup after an hour of user load run.
o After hour and half after the full backup and verify, scheduled transaction log backups
for every 30 minutes for seven hours
o Differential backup after seven and a half hours.
• The backups were verified for integrity and checksums by enabling
Verify backup when
finished
and Perform Checksum before writing to the media in the backup options screen.
• Performed a restore to a new database from the full, differential, and transaction log backups.
The restored database was verified using the T-SQL query,
DBCC CHECKDB .
For this test we performed full, log, and differential backups as shown in Figure 21. The performance
counters measured for this test were CPU utilization, application response times, IO operations at the
storage array, backup/restore times from the SQL server, and disk read latency at the storage array.
The sections below highlight the performance impact at each step (i.e. full backup, transaction log
backup, differential backup, and restore).
Figure 21 Backup operations in test #2