11gRAC CHM Management

Cluster Health Monitor (abbreviated CHM) is an Oracle-provided tool that uses its own proactive resources to collect operating systems (CPU, memory, SWAP, process, I/O, and Network) usage.

CHM data is collected once per second, 11.2.0.3 The version number is 5 seconds to collect.

These system resource data are helpful for diagnosing node restart, hang, instance eviction (eviction), performance problem, etc. of cluster system.

The use of CHM is also used to detect some problems such as high system load, memory anomalies and so on, so as to avoid more serious problems.

CHM will install the following software on its own initiative:

11.2.0.2 and later versions of Oracle gridinfrastructure for Linux (without Linux Itanium), Solaris (Sparc 64 and x86-64)

11.2.0.3 and later version Oracle gridinfrastructure for AIX, Windows (not including Windows Itanium).

In the cluster, you can view the status of the appropriate CHM resources (ORA.CRF) by using the following commands:

$ crsctl Stat res ora.crf-init

The CHM consists mainly of two services:

1). Systemmonitor Service (Osysmond): This service will be executed on all nodes. Osysmond will send the resource usage of each node to the cluster logger service, which will receive all node information and save it to the CHM repository.

2). Cluster Logger Service (ologgerd): In a cluster. Ologgerd will have a host point (master). Another standby node (standby). When Ologgerd fails to start after the current node encounters a problem, it is enabled on the standby node. (This and DRM's master are different concepts)

CHM Repository: Used to store collected data . By default. Will exist in the grid Infrastructure home, requires 1 GB of disk space, each node will occupy approximately 0.5GB of space per day. You can use Oclumon to adjust its storage path and the size of the agreed space (up to only 3 days of data).

Close and open CHM (preferably run on 2 nodes with a grid account)

Shut down:

ora11grac1

ora11grac2

Open:

ora11grac1

ora11grac2

Note:

1. I/O drops slowly after shutting down the service

2. This change only applies to this time, restart DB or CRS service This change is invalid

3. Whether to enable device IO that depends primarily on the production environment, the test machine can be directly disabled

To enable and disable CHM:

#<grid_home>/bin/crsctl Modify Resource Ora.crf-attr "Auto_start=never" –init

#<grid_home>/bin/crsctl Modify Resource Ora.crf-attr "Auto_start=always"-init

Reference Document: MoS file: Cluster health Monitor (CHM) FAQ (Doc ID 1328466.1)

11gRAC CHM Management