Common Scenarios to avoid with DataWarehousing

Database Design

 

Rule	Description	Value	Source	Problem Description
1	Excessive sorting and RID lookup operations   should be reduced with covered indexes.		Sys.dm_exec_sql_text Sys.dm_exec_cached_plans	Large data warehouse can benefit from more   indexes. Indexes can be used to cover queries and avoid sorting. The cost of   index overhead is only paid when data is loaded.
2	Excessive fragmentation: Average fragmentation_in_percent should be   <25%	>25%	sys.dm_db _index_physical_stats	Reducing index fragmentation through index   rebuilds can benefit big range scans, common in data warehouse and Reporting   scenarios.
3	Scans and ranges are common. Look for missing   indexes	>= 1	Perfmon object SQL Server Access Methods Sys.dm_db_missing_index_group_stats Sys.dm_db_missing_index_groups Sys.dm_db_missing_index_details	A missing index flushes the cache.
4	Unused Indexes should be avoided	If an index is NEVER used, it will not appear   in the DMV sys.dm_db_index_usage_stats		Index maintenance for unused indexes should be   avoided.

 

Resource issue: CPU

 

Rule	Description	Value	Source	Problem Description
1	Signal Waits	> 25%	Sys.dm_os_wait_stats	Time in runnable queue is pure CPU wait.
2	Avoid plan reuse	> 25%	Perfmon object SQL Server Statistics	Data warehouse has fewer transactions than   OLTP, each with significantly bigger IO. Therefore, having the correct plan   is more important than reusing a plan. Unlike OLTP, data warehouse queries   are not identical.
3	Parallelism: Cxpacket waits	<10%	Sys.dm_os_wait_stats	Parallelism is desirable in data warehouse or reporting   workloads.

 

Resource issue: Memory

 

Rule	Description	Value	Source	Problem Description
1	Memory grants pending	>1	Perfmon object SQL Server Memory Manager	Memory grant not available for query to run.  Check for Sufficient memory and page life expectancy.
2	Page life expectancy	Drops by 50%	Perfmon object SQL Server Buffer Manager	Page life expectancy is the average number of   seconds a data page stays in cache.    Low values could indicate a cache flush that is caused by a big read.  Look for possible missing index.

 

Resource issue: IO

 

Rule	Description	Value	Source	Problem Description
1	Average Disk sec/read	>20 ms	Perfmon object Physical Disk	Reads should take 4-8ms without any IO   pressure.
2	Average Disk sec/write	>20 ms	Perfmon object Physical Disk	Writes (sequential) can be as fast as 1 ms for   transaction log.
3	Big scans	>1	Perfmon object SQL Server Access Methods	A missing index flushes the cache.
4	If Top 2 values for wait stats are any of the   following: ASYNCH_IO_COMPLETION IO_COMPLETION LOGMGR WRITELOG PAGEIOLATCH_x	Top 2	Sys.dm_os_wait_stats	If top 2 wait_stats values include IO, there   is an IO bottleneck

 

Resource issue: Blocking

 

Rule	Description	Value	Source	Problem Description
1	Block percentage	>2%	Sys.dm_db_index_operational_stats	Frequency of blocks.
2	Block process report	30 sec	Sp_configure, profiler	Report of statements.
3	Average Row Lock Waits	>100ms	Sys.dm_db_index_operational_stats	Duration of blocks.
4	If Top 2 values for   wait stats are any of the following: LCK_M_BU LCK_M_IS LCK_M_IU LCK_M_IX LCK_M_RIn_NL LCK_M_RIn_S LCK_M_RIn_U LCK_M_RIn_X LCK_M_RS_S LCK_M_RS_U LCK_M_RX_S LCK_M_RX_U LCK_M_RX_X LCK_M_S LCK_M_SCH_M LCK_M_SCH_S LCK_M_SIU LCK_M_SIX LCK_M_U LCK_M_UIX LCK_M_X	Top 2	Sys.dm_os_wait_stats	If top 2 wait_stats   values include IO, there is a blocking bottleneck. Consider using row   versioning to minimize shared locking blocks.

 

Exactly the opposite of OLTP applications, reporting or relational data warehouse applications are characterized by small numbers of (different) big transactions. These are frequently SELECT intensive operations. The implications are significant for database design, resource usage, and system performance.

Reporting and data warehouse performance objectives are as follows:

1. Data warehouse and relational data warehouse designs can have more indexes as the cost of index maintenance is paid only one time, during the batch update process.
2. Plan reuse should generally be avoided. Plan reuse may result in picking up a plan that was good for some other query (with different data distribution), but may not be good for this query.  The time taken for plan generation of a large DataWarehouse query is not nearly as important as having the right plan. 
3. Sorts can and should be minimized with correct index usage.
4. Missing index situations should be investigated and corrected.
5. Large IOs such as range scans benefits from on disk contiguity. Index fragmentation should be frequently monitored and kept to a minimum with index rebuilds.
6. Blocking is generally uncommon as most data warehouse transactions are read operations.
7. Parallelism is generally desirable for data warehouse applications.