Thursday, 20 July 2017

Oracle wait event

    db file sequential reads (unselective/Fragmented  index)
        rebuild index
    db file scattered reads (Full Table scans/Fast Full Index Scans)
        indexes has been dropped or become unusable
    log file parallel write(LGWR waits redo log buffer cache to redo log)
    log file sync(FG waiting for a COMMIT or ROLLBACK)
        reduce commits
    buffer busy waits ( Another session reading or holding blocks in buffer )
        Data Blocks - check index, reduce number of rows per block
        Undo Header - increase Rollback Segments
    free buffer waits (buffer cache is too small/Checkpoints happening too fast)
        Reduce checkpoint frequency
    enqueue waits (wait for a lock)
        Reduce waits and wait times
        Recreate all temporary tablespaces
    Cache buffer chain latch
    Direct Path Reads (read data directly in pga/require sorts ORDER BY, GROUP BY, UNION, DISTINCT, and ROLLUP)
        Set PGA_AGGREGATE_TARGET to appropriate value
    Direct Path  Writes (write data from users’ PGAs to data files)
    Latch Free Waits(process is waiting for a latch that is currently busy)
    Library cache latch
    Shared pool latch (due to intense hard  parsing)
    Row cache objects latch ()
        increasing the size of the shared pool

No comments:

Post a Comment