Forgot your password?    
+ Reply to Thread
Results 1 to 2 of 2

Thread: 11gR2 RAC Wait Events

  1. #1
    Expert Oracle Administrator
    Join Date
    Oct 2011
    Location
    New Delhi, India
    Posts
    427

    11gR2 RAC Wait Events

    Wait class “Cluster” indicates cluster-related waits related to Cache Fusion:

    SQL> select wait_class from v$system_wait_class;

    WAIT_CLASS
    ----------------------------------------------------------------
    Other
    Application
    Configuration
    Concurrency
    Commit
    Idle
    Network
    User I/O
    System I/O
    Cluster


    RAC Wait Event Naming Conventions

    - Events that start with “GCS%” and “gc%” are Cache Fusion-related waits. In other word, they have to do with waits experienced in shipping current or consistent-read versions of blocks across instances in a RAC cluster
    - Events that start with “ges%’ are related to Global Enqueue Services


    gc current/cr multiblock request = is only a placeholder , until the block is obtained


    Block provided Immediate from global Cache

    - “gc current block 2-way” - Indicates that a current block was sent from either the resource master requiring 1 message and 1 transfer
    - “gc current block 3-way” – Indicates that a current block was forwarded to a 3rd node from which it was sent, requiring 2 messages and one block transfer.
    - “gc cr block 2-way” - Indicates that a cr block was sent from either the resource master requiring 1 message and 1 transfer
    - “gc cr block 3-way” -Indicates that a current block was forwarded to a 3rd node from which it was sent, requiring 2 messages and one block transfer



    Grant for Disk I/O RAC Waits

    - “gc current grant 2-way” – Indicates that no current block was received because it was not cached in any instance. Instead a global grant was given, enabling the requesting instance to read the block from disk.
    - “gc cr grant 2-way” - – Indicates that no current block was received because it was not cached in any instance. Instead a global grant was given, enabling the requesting instance to read the block from disk and build a consistent read image for the block.




    Block Contention RAC Waits

    - gc current block busy” – Indicates a local instance made a request for a current version of a block and did not immediately receive the block. This indicates that the block shipping was delayed on the remote instance. Could be on account of block contention.

    A buffer may also be busy locally when a session has already initiated a cache fusion operation and is waiting for its completion when another session on the same node is trying to read or modify the same data. High service times for blocks exchanged in the global cache may exacerbate the contention, which can be caused by frequent concurrent read and write accesses to the same data.

    - “gc cr block busy” – Indicates a local instance made a request for a CR version of a block and did not immediately receive the block

    - gc current grant busy” – Indicates a local instance made a request for a current version of a block and did not immediately receive the grant.
    This can occur because of multiple current block images / blocks being formatted on account of HWM extension.

    - “gc buffer busy acquire” – Indicates that the local instance cannot grant access to data in the local buffer cache because a global operation on the buffer is pending and not yet completed

    - “gc buffer busy release” – Indicates that a remote instance is waiting to complete access on a block to prep for shipping to a remote instance.

    - “gc buffer busy” events mean that there is block contention that is resulting from multiple local requests for the same block, and Oracle must queue these requests.

    - "gc current/cr failure/retry" signifies a hardware or a network problem.


    High Load wait Events

    - “gc current block congested”
    - “gc cr block congested”

    The load wait events indicate that a delay in processing has occurred in the GCS, which is usually caused by high load, CPU saturation,paging and would have to be solved by additional CPUs, load-balancing, off loading processing to different times or a new cluster node.



    AWR Reports

    Check the values for:
    Average global cache cr block recieve time (ms)
    Average global cache current block recieve time (ms)

    These statistics should not get worse over time.

  2. #2
    Oracle Administrator
    Join Date
    Dec 2011
    Posts
    92
    Some more wait events.

    1) Global cache blocks lost- This statistic shows block losses during network
    transfers. High values of this statistic indicate network problems.
    The use of an unreliable IPC protocol, such as UDP, may result in the value
    for lost global cache blocks being non-zero.

    High values for this statistics indicates that there is some problem with the IPC,
    network or hardware.

    2) Global cache blocks corrupt - This statistic shows if there is any block corruption
    during interconnect transfers. High value for this statistic indicates
    that there is a IPC,network or hardware problem.

    3) global cache open s and global cache open x - These events are generated when
    an instance initialy fetches data blocks. The duration of the wait event should be
    short.

    This wait event is generated when the blocks being requested are not in the
    cache of any instance in the cluster, which in turn results into reads from disk.

    Another reason for this wait event can be that the blocks cannot be fetched
    from other instances cache and becuase of that there is a disk read.

    One way to reduce this wait event is to preload the buffer cache with frequently
    accessed tables.


    4) global cache null to s and global cache null to x - These wait events are
    generated whenever there is a inter-instance block ping across the network.

    Inter-instance block ping occurs when two instances exchange the
    same block between each other repeatedly.

    Processes waiting for global cache null to s events are waiting for
    a block to be transferred from the instance that last changed it.

    This event consumes a large proportion of wait time, When one instance
    frequently requests cached data blocks from the other RAC instances.

    Recommended method to reduce this wait event is to reduce number of rows
    per block to reduce block swapping among nodes.

    5) Global cache cr request - This event is generated when an instance requested a
    consistent read data block and the block to be transferred has not arrived to
    the instance who had requested it.

    6) gc cr block lost - Possible reasons for this event are as following:

    i) Network problems(discarded packets and fragments, packet reassembly or timeouts,
    buffer overflows etc).

    ii) Checksum errors or corrupted headers.

    Possible solution is to check our entire network layer for problems ex NIC,switch etc.


    7) gc buffer busy - GC buffer busy means that the buffer in the buffer cache, that the
    session is trying to access is already involved in another ongoing global cache operation.
    Until that global cache operation completes, session must wait.
    Possible reasons for this event may be due to CPU starvation issues, Swapping issues, interconnect
    issues etc.

    8) congested - The events that contain "congested" suggest CPU saturation
    (runaway or spinning processes), long running queues, and network
    configuration issues. Excessive swapping in harddisk because of low memory can
    also cause this event.

    9) busy - The events that contain "busy" indicate contention. Investigation for
    this event should be done by examining SQL with high cluster wait time or segments
    with high interconnect block transfers.

    10) Gc[current/cr][failure/retry] - This event means that due to network problems,
    the sending instance needs to retry sending the block image.

    This event happen when there is network or hardware problems.

    11) Gc [current/cr] [2/3]-way - If we have a two-node cluster, then only a 2-way
    cluster is possible, as we can have at most two hops. If we have three or more
    RAC instances, then 2-way or 3-way cluster is possible. Events are received
    immediately after two or three network hops.

    12) Gc[current/cr] grant 2-way - A grant means, that an instance has requested
    a block image and no other instance in the cluster has it in its buffer cache.

    So, the requesting instance performs a disk i/o and gets the data from the
    datafile.

    A grant can be of two types "CR" or "Current".

    A current grant reads the block from the datafiles.

    A cr grant reads the block from disk and builds a read consistent block,
    as of a certain time in past.

    13) Gc[current/cr][block/grant] congested - A current or cr block is requested and a block or
    grant message is received. The congested hint implies that the request spent more than 1 ms in internal
    queues. Reasons for this event may include intensive CPU consumption, lack of memory,
    excessive paging and swapping.

    14) Gc[current/cr] block busy - A current or cr block is requested and received,
    but is not sent immediately by LMS which caused the delay in sending.

    15) Gc current grant busy - A current block is requested and a grant message received.
    The busy hint means that the request is blocked for some reason or it
    cannot be handled immediately.


    16) Gc[current/cr][failure/retry] - A block is requested and a failure status
    received or some other exceptional event has occured.


    17) Gc buffer busy - If the time between buffer accesses becomes less than the time the
    buffer is pinned in memory, the buffer containing a block is said to become busy and as a result
    interested users may have to wait for it to be unpinned.

+ Reply to Thread

Similar Threads

  1. Wait Events
    By devesh_ocp in forum Core Database Administration and Monitoring
    Replies: 0
    Last Post: 04-25-2012, 12:01 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts

DBA Lounge (P) Ltd. deals in Oracle Technologies on Consulting, Resourcing, Corporate Training


Online and corporate training available on Oracle Database 11g, Oracle 11g Real Application Cluster (RAC), Oracle Applications 11i/R12, Oracle Fusion Middleware 11g, Oracle Identity Management-OIM, Oracle Internet Directory 11g-OID, Oracle Business Intelligence Enterprise Edition-OBIEE, Oracle Golden Gate, Oracle Access Management-OAM, Oracle Internet Directory-ODS, Oracle Identity Analytics Architecture-OIA Statistics