Tuesday, May 1, 2012

Performance Tuning very common Oracle wait events

As there are over 1000+ wait events but you may come across below mentioned few wait events very frequently.
In this post I try to cover most popular of them.

To get the list of all events you can query V$EVENT_NAME.

db file scattered reads
Possible Causes :
· The Oracle session has requested and is waiting for multiple contiguous database blocks (up to DB_FILE_MULTIBLOCK_READ_COUNT)  to be  read into the SGA from disk.
· Full Table scans
· Fast Full Index Scans
Actions :
· Optimize multi-block I/O by setting the parameter DB_FILE_MULTIBLOCK_READ_COUNT
· Partition pruning to reduce number of blocks visited
· Consider the usage of multiple buffer pools and cache frequently used indexes/tables in the KEEP pool
· Optimize the SQL statement that initiated most of the waits. The goal is to minimize the number of physical
and logical reads.
· Should the statement access the data by a full table scan or index FFS? Would an index range or unique scan
 be more efficient? Does the query use the right driving table?
· Are the SQL predicates appropriate for hash or merge join?
· If full scans are appropriate, can  parallel query improve the response time?
· The objective is to reduce the demands for both the logical and physical I/Os, and this is best
achieved through SQL and application tuning.
· Make sure all statistics are representative of the actual data. Check the LAST_ANALYZED date
Remarks:
· If an application that has been running fine for a while suddenly clocks a lot of time on the db file scattered read event and there hasn’t been a code change, you might want to check to see if one or more indexes has been dropped or become unusable.
· Or  whether the stats has been stale.
db file sequential reads
Possible Causes :
· Use of an unselective index 
· Fragmented Indexes
· High I/O on a particular disk or mount point
· Bad application design 
· Index reads performance can be affected by  slow I/O subsystem and/or poor database  files layout, which result in a higher average  wait time
Actions :
· Check indexes on the table to ensure that the right index is being used
· Check the column order of the index  with the WHERE clause of the Top SQL statements
· Rebuild indexes with a high clustering factor
· Use partitioning to reduce the amount of blocks being visited
· Make sure optimizer statistics are up to date
· Relocate ‘hot’ datafiles
· Consider the usage of multiple buffer pools and cache frequently used indexes/tables in the KEEP pool
· Inspect the execution plans of the SQL statements that access data through indexes
· Is it appropriate for the SQL statements to access data through index lookups?
· Would full table scans be more efficient?
· Do the statements use the right driving  table?
· The optimization goal is to minimize  both the number of logical and physical I/Os.
Remarks:
· The Oracle process wants a block that is currently not in the SGA, and it is waiting for the database block to be read into the SGA from disk.
· Significant db file sequential read wait time is most likely an application issue.
· If the DBA_INDEXES.CLUSTERING_FACTOR of the index approaches the number of blocks in the table, then most of the rows in the table are ordered. This is desirable.
· However, if the clustering factor approaches the number of rows in the table, it means the rows in the table are randomly ordered and thus it requires more I/Os to complete the operation. You can improve the index’s clustering factor by rebuilding the table so that rows are ordered according to the index key and rebuilding the index thereafter.
· The OPTIMIZER_INDEX_COST_ADJ and OPTIMIZER_INDEX_CACHING initialization parameters can influence the optimizer to favour the nested loops operation and choose an index access path over a full table scan.
log file parallel write
Possible Causes :
· LGWR waits while writing contents of the redo log buffer cache to the online log files on disk
· I/O wait on sub system holding the online  redo log files
Actions :
· Reduce the amount of redo being generated
· Do not leave tablespaces in hot backup mode for longer than necessary
· Do not use RAID 5 for redo log files
· Use faster disks for redo log files
· Ensure that the disks holding the archived redo log files and the online redo log files are separate so as to avoid contention
· Consider using NOLOGGING or UNRECOVERABLE options in SQL statements

log file sync:

Possible Causes :
· Oracle foreground processes are waiting for a COMMIT or ROLLBACK to complete
Actions :
· Tune LGWR to get good throughput to  disk eg: Do not put redo logs on  RAID5
· Reduce overall number of commits by batching transactions so that there are fewer distinct COMMIT operations
Actions :
  • Tune LGWR to get good throughput to  disk eg: Do not put redo logs on RAID5
  •  Reduce overall number of commits by batching transactions so that there are fewer distinct COMMIT operations
buffer busy waits:
Possible Causes :
· Buffer busy waits are common in an I/O-bound Oracle system.
· The two main cases where this can occur are:
· Another session is reading the block into the buffer
· Another session holds the buffer in an incompatible mode to our request
· These waits indicate read/read, read/write, or write/write contention.
· The Oracle session is waiting to pin a buffer .A buffer must be pinned before it can be read or modified. Only one process can pin a
buffer at any one time.
· This wait can be intensified by a large block  size as more rows can be contained within the block
· This wait happens when a session wants to access a database block in the buffer cache but it cannot as the buffer is “busy
· It is also often due to several processes repeatedly reading the same blocks (eg: i lots of people scan the same index or data block)
Actions :
· The main way to reduce buffer busy waits is to reduce the total I/O on the system
· Depending on the block type, the actions will differ
Data Blocks
· Eliminate HOT blocks from the application. Check for repeatedly scanned / unselective indexes.
· Try rebuilding the object with a higher PCTFREE so that you reduce the number of rows per block.
· 
 Check for ‘right- hand-indexes’ (indexes that get inserted into at the same point by many processes).
· Increase INITRANS and MAXTRANS and reduce PCTUSED This will make the table less dense .
· Reduce the number of rows per block
Segment Header
· Increase of number of FREELISTs   and FREELIST GROUPs
Undo Header
· Increase the number of Rollback Segments
free buffer waits:
Possible Causes :
· This means we are waiting for a free buffer but there are none available in the cache because there are too many dirty buffers in  the cache
· Either the buffer cache is too small or the DBWR is slow in writing modified buffers to disk
· DBWR is unable to keep up to the write  requests
· Checkpoints happening too fast – maybe due  to high database activity and under-sized  online redo log files
· Large sorts and full table scans are filling the cache with modified blocks faster than the  DBWR is able to write to disk
· If the  number of dirty buffers that need to be  written to disk is larger than the number that DBWR can write per batch, then these waits  can be observed
Actions :
Reduce checkpoint frequency  – increase the size of the online redo log files
Examine the size of the buffer cache – consider increasing the size of the buffer cache in the SGA
Set disk_asynch_io = true set
If not using asynchronous I/O increase the number of db writer processes or dbwr slaves
Ensure hot spots do not exist by spreading datafiles over disks and disk controllers
Pre-sorting or reorganizing data can help
enqueue waits
Possible Causes :
· This wait event indicates a wait for a lock  that is held by another session (or sessions) in an incompatible mode to the requested mode.
 TX Transaction Lock
· Generally due to table or application set up issues
· This indicates contention for row-level lock. This wait occurs when a transaction tries to update or delete rows that are currently
 locked by another transaction.
· This usually is an application issue.
TM DML enqueue lock
· Generally due to application issues, particularly if foreign key constraints have not been indexed.
ST lock
· Database actions that modify the UET$ (used extent) and FET$ (free extent) tables require the ST lock, which includes actions such as drop, truncate, and coalesce.
· Contention for the ST lock indicates there are multiple sessions actively performing
· dynamic disk space allocation or deallocation
· in dictionary managed tablespaces
Actions :
· Reduce waits and wait times
· The action to take depends on the lock  type which is causing the most problems
· Whenever you see an enqueue wait event for the TX enqueue, the first step is to find out who the blocker is and if there are multiple waiters for the same resource
· Waits for TM enqueue in Mode 3 are primarily due to unindexed foreign key columns.
· Create indexes on foreign keys  < 10g
· Following are some of the things you can do to minimize ST lock contention in your database:
· Use locally managed tablespaces
· Recreate all temporary tablespaces  using the CREATE TEMPORARY TABLESPACE TEMPFILE… command.
Cache buffer chain latch
Possible Causes :
· Processes need to get this latch when they  need to move buffers based on the LRU block replacement policy in the buffer cache
· The cache buffer lru chain latch is acquired in order to introduce a new block into the buffer cache and when writing a buffer
back to disk, specifically when trying  to scan the LRU (least recently used) chain containing all the dirty blocks in the buffer
cache. Competition for the cache buffers lru chain .
· latch is symptomatic of intense buffer cache  activity caused by inefficient SQL  statements. Statements that repeatedly scan
· large unselective indexes or perform full table scans are the prime culprits. 
· Heavy contention for this latch is generally  due to heavy buffer cache activity which  can be caused, for example, by:
 Repeatedly scanning large unselective indexes
Actions :
 Contention in this latch can be avoided implementing multiple buffer pools or increasing the number of LRU latches with the  parameter DB_BLOCK_LRU_LATCHES (The default value is generally  sufficient for most systems).
Its possible to reduce contention for the cache buffer lru chain latch by increasing the  size of the buffer cache and  thereby reducing the rate at which new blocks are  introduced into the buffer cache.
 Direct Path Reads
Possible Causes :
· These waits are associated with direct read operations which read data directly into the sessions PGA bypassing the SGA
· The “direct path read” and “direct path write” wait events are related to operations that are performed in PGA like sorting, group by operation, hash join
· In DSS type systems, or during heavy batch periods, waits on “direct path read” are quite normal However, for an OLTP system these waits are significant
· These wait events can occur during sorting operations which is not surprising as direct path reads and writes usually occur in connection with temporary tsegments
· SQL statements with functions that require sorts, such as ORDER BY, GROUP BY, UNION, DISTINCT, and ROLLUP, write sort runs to the temporary tablespace when the input size is larger than the work area in the PGA
Actions :
Ensure the OS asynchronous IO is configured correctly.
Check for IO heavy sessions / SQL and see if the amount of IO can be reduced.
Ensure no disks are IO bound.
Set your PGA_AGGREGATE_TARGET to appropriate value (if the parameter WORKAREA_SIZE_POLICY = AUTO) Or set *_area_size manually (like sort_area_size and then you have to set WORKAREA_SIZE_POLICY = MANUAL
Whenever possible use UNION ALL instead of UNION, and where applicable use HASH JOIN instead of SORT MERGE and NESTED LOOPS instead of HASH JOIN.
 Make sure the optimizer selects the right driving table. Check to see if the composite index’s columns can be rearranged to match the ORDER BY clause to avoid sort entirely.
Also, consider automating the SQL work areas using PGA_AGGREGATE_TARGET in Oracle9i Database.
Query V$SESSTAT> to identify sessions with high “physical reads direct”
Remark:
· Default size of HASH_AREA_SIZE  is twice that of SORT_AREA_SIZE
· Larger HASH_AREA_SIZE will influence optimizer to go for hash joins instead of nested loops
· Hidden parameter DB_FILE_DIRECT_IO_COUNT can impact the direct path read performance.It sets the maximum I/O buffer size of direct read and write operations. Default is 1M in 9i
Direct Path  Writes:
Possible Causes :
· These are waits that are associated with direct write operations that write data from users’ PGAs to data files or temporary  tablespaces
· Direct load operations (eg: Create Table as  Select (CTAS) may use this)
· Parallel DML operations
· Sort IO (when a sort does not fit in memory
Actions :
If the file indicates a temporary  tablespace check for unexpected disk sort operations.
Ensure
<Parameter:DISK_ASYNCH_IO> is TRUE . This is unlikely to reduce wait times from the wait event timings but
may reduce sessions elapsed times (as synchronous direct IO is not accounted for in wait event timings).
Ensure the OS asynchronous IO is configured correctly.
Ensure no disks are IO bound
Latch Free Waits
Possible Causes :
· This wait indicates that the process is waiting for a latch that is currently busy  (held by another process).
· When you see a latch free wait event in the V$SESSION_WAIT view, it means the process failed to obtain the latch in the
willing-to-wait mode after spinning  _SPIN_COUNT times and went to sleep. When processes compete heavily for  latches, they will also consume more CPU resources because of spinning. The result is a higher response time
Actions :
· If the TIME spent waiting for latches is significant then it is best to determine which latches are suffering from contention.
Remark:
· A latch is a kind of low level lock. Latches apply only to memory structures in the SGA. They do not apply to database objects. An Oracle SGA has many latches, and they exist to protect various memory structures from potential corruption  by concurrent access.
· The time spent on latch waits is an effect, not a cause; the cause is that you are doing too many block gets, and block gets require cache buffer chain latching
 Library cache latch
Possible Causes :
· The library cache latches protect the  cached SQL statements and objects definitions held in the library cache within the shared pool. The library cache latch must be acquired in order to add a new statement to the library cache.
· Application is making heavy use of literal SQL- use of bind variables will reduce this latch considerably
Actions :
· Latch is to ensure that the application is reusing as much as possible SQL statement representation. Use bind variables whenever ossible in the application.
· You can reduce the library cache latch hold time by properly setting the SESSION_CACHED_CURSORS parameter.
· Consider increasing shared pool.
Remark:
· Larger shared pools tend to have  long free lists and processes that need to allocate space in them must  spend extra time scanning the long free lists while holding the shared pool latch
· if your database is not yet on  Oracle9i Database, an oversized shared pool can increase the contention for the shared pool latch..
Shared pool latch
Possible Causes :
The shared pool latch is used to protect critical operations when allocating and freeing memory in the shared pool
Contentions for the shared pool and library cache latches are mainly due to intense hard  parsing. A hard parse applies to new cursors and cursors that are aged out and must be re-executed
The cost of parsing a new SQL statement is expensive both in terms of CPU requirements and the number of times  the library cache and shared pool latches  may need to be acquired and released.
Actions :
· Ways to reduce the shared pool latch  are, avoid hard parses when possible, parse once, execute many.
· Eliminating literal SQL is also useful to avoid the shared pool latch. The size  of the shared_pool and use of MTS  (shared server option) also greatly  influences the shared pool latch.
· The workaround is to set the initialization parameter  CURSOR_SHARING to FORCE. This  allows statements that differ in literal
 values but are otherwise identical to share a cursor and therefore reduce latch contention, memory usage, and  hard parse.
Row cache objects latch
Possible Causes :
This latch comes into play when user processes are attempting to  access the cached data dictionary values.
Actions :
· It is not common to have contention in this latch and the only way to reduce contention for this latch is by increasing the size of the shared pool (SHARED_POOL_SIZE).
· Use Locally Managed tablespaces for your application objects especially indexes 
· Review and amend your database logical design , a good example is to merge or decrease the number of  indexes on tables with heavy inserts
Remark:
· Configuring the library cache to an acceptable size usually ensures that the data  dictionary cache is also properly sized. So tuning Library Cache will tune Row Cache indirectly.