Home > Workload Solutions > SQL Server > White Papers > Dell PowerMax 2500 and 8500 Best Practices for Mission Critical SQL Server Databases > PowerMax data reduction
The PowerMax storage system provides data reduction without compromising performance. The PowerMax data reduction is the combination of the following components:
Also, note that:
For more information about PowerMax data reduction, see Dell PowerMax: Data Reduction inline compression and deduplication
While PowerMax local replications such as snapshots and clones natively benefit from efficient data reduction due to shared storage allocations and thin provisioning, when host-based copies are made, PowerMax deduplication comes into effect. Deduplication is especially effective for data copies that are aligned to the PowerMax track size, which is 128 KB. Therefore, it is recommended to format NTFS file systems with a 128 KB allocation unit.
In the following example, a Windows file copy is performed between two drives formatted with NTFS with a 128 KB allocation unit. Each drive is based on a PowerMax volume from an SG that has data reduction enabled.
As seen in the preceding figure, drive D:\ has three user database directories with a total of 329 GB used space. Drive F:\ is 100 percent free (freshly formatted disk with 128 KB allocation units).
If we look at the storage metrics before the copy operation using the Solutions Enabler CLI, we see the SRP capacity showing 1637.7 GB used. The source volume reports 296.2 GB effective capacity used, but only 60.1 GB physical capacity is actually taken. As a result, the Data Reduction Ratio (DRR) is 4.9:1. The target volume only has 0.1 GB effective capacity used, so we can ignore its DRR until after the copy.
# symcfg list -srp -detail -gb
STORAGE RESOURCE POOLS
Symmetrix ID : 000120200305
C A P A C I T Y
-------------------------------- --- --------------------------------------------
Flg Physical Physical Physical Provisioned
Name DRR (GB) Used (GB) Free (GB) (GB)
-------------------------------- --- ---------- ---------- ---------- -----------
SRP_1 BXO 74682.7 1637.7 73045.0 24334.5
---------- ---------- ---------- -----------
Total 74682.7 1637.7 73045.0 24334.5
# symcfg list -sg_compression -all | egrep 'Flags|Devices|DRR'
Flags Number Effective Physical Estimated
Storage Group Name C Devices Used (GB) Used (GB) Ratio
WinDRR1 X 1 296.2 60.1 1.7:1
WINDRR2 X 1 0.1 0.0 26.9:1
# symcfg list -srp -demand -type sg | egrep 'Effective|Used|DRR'
Effective Snapshot Snapshot Data
Provisioned Used Effective Resources Reduction
SG Name (GB) (GB) (%) Used (GB) Used (%) Ratio
WinDRR1 1024.0 296.2 28 0.0 0.0 4.9:1
WINDRR2 1024.0 0.1 0 0.0 0.0 179.9:1
Next, we copy the three user database directories in Windows File Explorer from drive D:\ to drive F:\. When the copy is finished, we can use the Solutions Enabler CLI to view the storage metrics.
The SRP-used capacity increased by 128.4 GB (1766.1 – 1637.7). The target volume shows 291.5 GB effective capacity used and 139.3 GB physical capacity used, with only 2.1:1 DRR.
# symcfg list -srp -detail -gb
STORAGE RESOURCE POOLS
Symmetrix ID : 000120200305
C A P A C I T Y
-------------------------------- --- --------------------------------------------
Flg Physical Physical Physical Provisioned
Name DRR (GB) Used (GB) Free (GB) (GB)
-------------------------------- --- ---------- ---------- ---------- -----------
SRP_1 BXO 74682.7 1766.1 72916.6 24334.4
---------- ---------- ---------- -----------
Total 74682.7 1766.1 72916.6 24334.4
# symcfg list -sg_compression -all | egrep 'Flags|Devices|DRR'
Flags Number Effective Physical Estimated
Storage Group Name C Devices Used (GB) Used (GB) Ratio
WinDRR1 X 1 296.2 49.7 1.7:1
WINDRR2 X 1 291.5 139.3 1.7:1
# symcfg list -srp -demand -type sg | egrep 'Effective|Used|DRR'
Effective Snapshot Snapshot Data
Provisioned Used Effective Resources Reduction
SG Name (GB) (GB) (%) Used (GB) Used (%) Ratio
WinDRR1 1024.0 296.2 28 0.0 0.0 6.0:1
WINDRR2 1024.0 291.5 28 0.0 0.0 2.1:1
Why did we not get an immediate 100 percent deduplication rate?
There are two reasons:
Typically, customer data is more compressible and there is ample other activity in the system besides copying three directories. In this particular test environment, we wait 30 minutes to view the metrics again, as shown in the following example:
# symcfg list -srp -detail -gb
STORAGE RESOURCE POOLS
Symmetrix ID : 000120200305
C A P A C I T Y
-------------------------------- --- --------------------------------------------
Flg Physical Physical Physical Provisioned
Name DRR (GB) Used (GB) Free (GB) (GB)
-------------------------------- --- ---------- ---------- ---------- -----------
SRP_1 BXO 74682.7 1637.7 73045.0 24334.5
---------- ---------- ---------- -----------
Total 74682.7 1637.7 73045.0 24334.5
# symcfg list -sg_compression -all | egrep 'Flags|Devices|DRR'
Flags Number Effective Physical Estimated
Storage Group Name C Devices Used (GB) Used (GB) Ratio
WinDRR1 X 1 296.2 31.1 1.7:1
WINDRR2 X 1 291.5 29.0 1.7:1
# symcfg list -srp -demand -type sg | egrep 'Effective|Used|DRR'
Effective Snapshot Snapshot Data
Provisioned Used Effective Resources Reduction
SG Name (GB) (GB) (%) Used (GB) Used (%) Ratio
WinDRR1 1024.0 296.2 28 0.0 0.0 9.5:1
WINDRR2 1024.0 291.5 28 0.0 0.0 10.1:1
We see that the SRP-used capacity went back to 1637.6 GB, which is the same capacity as before the copy and demonstrates a 100 percent deduplication rate.
We also see that the physical capacity of both source and target devices is much lower compared to the effective used capacity, resulting in DRR of 9.5:1 (296.2 / 31.1) and 10.1:1 (291.5 / 29.0).