Shulou(Shulou.com)06/01 Report--

Details of the rac process:

1. RAC background process

LMON:LOCK Monitor Processes is also known as Global enqueue service monitor

Monitor the status of the entire cluster and maintain the memory structure of GCS

Monitor processes and instances that are abnormally terminated

Reconfiguration of locks and resources when the instance leaves and joins the cluster

Manage global locks and resources

Monitor global lock resources, handle deadlocks and blocking

LMSn:Lock Monitor Services is also known as GCS (Global Cache Services) processes

The LMS process is mainly used to manage database access within the cluster and to transfer block mirrors in the buffer cache of different instances, when on a data block

When a consistent read occurs, LMS is responsible for rolling back the block and copy it to the requested instance

Each RAC node has at least 2 LMS processes

LCK:Lock Process

The LCK process is mainly used to manage inter-instance resource requests and cross-instance invocation operations, including object access such as data dictionaries, and handle non-

CHACE resource request for CACEH FUSION (for example, dictionary cache or row cache request)

Because the LMS process is responsible for the main lock management functions, there is only one LCK process per instance

LMD:Lock Monitor Deamon Process

The LMD process mainly manages access to global queues and resources, updates the corresponding queue status, and processes resources from other instances. The current state of each global queue is stored in the corresponding instance shared memory. This status indicates that the instance has the corresponding right to use the resource. There is a special queue in the shared memory of an instance master that records resource requests from other remote instances. When the LMD process of the remote instance issues a resource request, the request points to the LMD of the master instance. When the LMD process of the master instance receives the request, it monitors whether the resource is invalid in a special queue in the shared memory. If the valid LMD process updates the status of the resource queue and notifies the LMD process requesting the resource that the resource queue is available, if the resource queue is being used by another instance or is currently invalid The LMD process notifies the LMD process application of the instance in use to release the resource, and when the resource release becomes valid, the LMD process of the master instance updates the status of the resource queue and notifies the LMD process requesting the resource instance that the resource queue is ready for use

DIAG:Diagnostic Deamon

New background process for oracle10g

Routinely monitor the health of the instance, as well as whether the instance hangs or has a deadlock

Collect critical diagnostic information in case of instance and process errors

This process updates the alert log file to write some important alarm information

II. RAC service process

CRS- Cluster Resource Service (cluster ready services)

Basic procedures for managing highly available operations in a cluster

Any transaction managed by CRS is called a resource

Database, instance, snooping, virtual IP, application process, etc.

CRS manages these resources according to the resource configuration information stored in OCR.

When the state of a resource changes, the CRS process generates an event

CSS- Cluster synchronization Service (Cluster Synchronization Service)

Manage the membership of cluster nodes

Controls which node is a member of the cluster, and the node notifies the cluster member when it joins or leaves the cluster to control the cluster configuration information

The failure of this process caused the cluster to restart

EVMD event Management Service (Event Management)

Event management daemon

The background process for publishing CRS creation events

Publish and subscribe Service for ONS- events (Oracle Notification Service)

Rapid application of communication notification event publishing and subscription service

OCR- Oracle Cluster Register

Cluster registration file, recording the relevant information of each node

Save all kinds of resource information of RAC cluster

Similar to the windows registry

Stored on a shared disk and shared by all instances

There are 2 alternate disks by default

Voting Disk voting disk

The arbitration mechanism is used to arbitrate the behavior written by multiple nodes to the shared node to avoid conflicts.

Stored on a shared disk and shared by all instances

Used to determine the relationship between individual instances

When a node fails, voting disk is used to decide which instance to expel

There are 3 alternate disks by default

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.