Oracle Clusterware Diagnosis - Expert Oracle RAC Performance Diagnostics and Tuning - page 542

High Availability Service (HAS)
Pre-Oracle 11g, the RAC clusterware stack consisted primarily of three daemon processes: the cluster synchronization
services (CSS), cluster ready services (CRS), and the event manager (EVM). Starting with Oracle Clusterware 11g
Release 2, the stack has been streamlined to two primary stacks: the upper stack, called the Cluster Ready Services
Daemon (CRSD), and the lower-level stack, called the Oracle High Availability Services Daemon (OHASD).
As illustrated in Figure 16-2 , OHASD is the primary cluster boot process when a server that contains Oracle
clusterware is started; it gets the cluster stack components started. While there are no real built-in levels within the
clusterware stack, for easy understanding the entire startup process is grouped into four levels in Figure 16-2 . Level 0
contains the clusterware boot process handled by OHASD.

1 2

0

3

4b

GNS
ora.gns

cssd
ora.cssd

ACFSRegistry
Ora.registry.acfs
Scan VIP
Ora.scan<d>.vip

crsd
ora.crsd

ctssd
ora.ctssd

gns.ip
Ora.gns.vip

diskmon
ora.diskmon

Node VIP
Ora.nodename.vip

acfs drivers
ora.drivers.acfs

Db
ora.DB.db

haip
ora.cluster_inerco
nnect.haip

Diskgroup
ora.dg

mdnsd
ora.mdnsd

ASM
ora.asm

gipcd
ora.gipcd

Network
ora.net<id>.network

gpnpd
ora.gpnpd

ONS
ora.ons

evmd
ora.evemd

Services
ora.<database>.svc

crf
ora.crf

GSD
ora.gsd

asm
ora.asm Listener

ora.listener

cssdmonitor
ora.cssdmonitor

scanlistener
ora.LISTENER_SCAN<id>.lsnr

4a

Figure 16-2. Clusterware boot-up process flow 3

3 Sandesh Rao and Bob Caldwell, “Troubleshooting and Diagnosing RAC and GI,” www.oracleracsig.org .