Clusterware startup sequence-OHASD

Clusterware startup sequence

[root@ebsdb1 etc] # crsctl check crs

CRS-4638: Oracle High Availability Services is online

CRS-4537: Cluster Ready Services is online

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

According to the above output, the cluster can be divided into four levels:

Level 1: OHAS level, responsible for initializing resources and processes of the cluster.

Level 2: CSS level, responsible for building the cluster and ensuring the consistency of the cluster.

Level 3: CRS level, responsible for managing various application resources of the cluster.

Level 4: the EVM level, which is responsible for passing cluster events between cluster nodes.

Next, the startup process at each level is described in detail:

OHASD level

This level is mainly responsible for initializing resources and processes of the cluster. The specific process is as follows:

The following line in 1./etc/inittab is called

$cat / etc/inittab | grep init.d | grep-v grep

H2:35:respawn:/etc/init.d/init.ohasd run > / dev/null 2 > & 1-net

< 私网编号 >

: net_ada=

< 私网网卡名 >

For example:

Gpnptool edit-p=$GPNPDIR/p.xml-o=$GPNPDIR/p.xml-ovr-prf_sq=9-net2:net_ada=eth2

6) confirm previous changes

Gpnptool sign-p=$GPNPDIR/p.xml-o=$GPNPDIR/p.xml-ovr-w=cw-fs:peer

7) apply the modified gpnp profile to the gpnpd daemon.

Gpnptool put-p=$GPNPDIR/p.xml

8) push the changed gpnp profile to other nodes in the cluster

Gpnptool find-c =

< 集群名 >

Gpnptool rget-c =

< 集群名 >

Method 3: confirm that the private network communication is normal (for example, use ping, traceroute and other commands to confirm the connectivity of the cluster private network)

Method 4: restart the gpnp daemon at the operating system level, for example: kill-9

When the gpnpd daemon is summarized, the corresponding ohasd agent process oraagent will discover the situation in time and start a new gpnpd daemon.

Method 5: confirm that the communication of the cluster private network is normal (for example, use ping, traceroute and other commands to confirm the connectivity of the cluster private network)

Method 6: restart the gipcd daemon at the operating system level, for example: kill-9

When the gipcd daemon is summarized, the corresponding ohasd agent process oraagent will discover the situation in time and start a new gipcd daemon.

Method 7: restart GI to rebuild the socket file.

Crsctl stop crs

Crsctl start crs