Oracle Rac Startup sequence

  Hello Friend's ,


In this post we will discuss about the Oracle Rac Startup sequence .

Its one of the favorite questions of interviewer for candidate with experience of over 3 years

So lets get started : 

There are 2 ways we can dopt it as short and detail version .

Short startup sequence :

1) OS

2) Cluster

3) ASM

4) INSTANCE

5) Listener

6) Service 

Detail explained is as below :

There are 4 Layers to the architecture of oracle as shown in below diagram 




 Credit source for diagram Internet .

Or : Simple One 


Now lets get started with the demons by their startup sequence:

Demons : 

1) INIT : INIT is the very first file to be checked to  be start  at the time of the operating system boot,it  is system and service manager for the Linux , Aix or any other Linux flavor. This service start and stop through script which is stored in the /etc/init.d directory .


2) CRSD : The CRS daemon (crsd) manage cluster resource based on the configuration information stored in the OCR for each resource.

Is responsible for the start,stop  monitor and failover of resource it maintain OCR and also restart the resource when the failover occurs ,

this applicable for RAC system . For standalone configuration OHASD is  used

Sequence : 

INIT --> INIT.OHASD --> OHASD --> CRSD [REBOOT]


3) Cluster synchronization service [CSS] : 


I) CSSD monitor -- Mandate : Monitor for the node hang & subsequently act as oprocd - which act as I/O fencing  and also monitors Ocssd [ via oclsomon] . This is a multithread process and runs with an elevated priority 

Sequence : 

INIT --> INIT.OHASD --> OHASD --> OHASD.bin --> CSSD monitor


II) CSSD Agent : This process is spawned by ohasd process responsible for  I/O fencing

Killing of process would cause node reboot , Also check the status ocssd.bin demon


Sequence : 

INIT --> INIT.OHASD --> OHASD --> OHASD.bin --> CSSD agent --> OCSSD


III) OCSSD.bin : This process manages the cluster node membership , failure of this process can cause node restart 

Sequence : 

INIT --> INIT.OHASD --> OHASD --> OHASD.bin --> CSSD agent --> OCSSD --> OCSSD.bin


4) EVMD  : 

I) EVMD.bin

II) EVMlogger.bin

III) EVMlogger.log


I) EVMD.bin : Distributes & communicates cluster events to all of the cluster member so that they are aware of the changes being done in the cluster 


II) EVMlogger.bin : Started by the evmd.bin reads the config file and determine the events to subscribe from evmd and it runs user defined reactions to those events 

 

5) Ora root agent : This is a specialized oraagent process which helps CRSD manage resouces owner by root e.g : Network , Node VIP , SCAN IP


  6) CTSSD :  The cluster time synchronization service process provides time synchronization for the cluster in the absence in the ntpd . if ntpd is configured , ctssd will run in observer mode.


7)  Cluster Logger Service : This  Logger receives information from all nodes in the cluster & persist CHM repository based on the database this service runs on only 2 nodes in cluster 

Cluster logger service  receives the os metric collection 

information from all the nodes and kept it in the Cluster  Health Monitor (CHM) repository , 

which received from  the osysmond  service.  These is only one cluster logger service (loggerd) 

run on entire cluster  any other node is selected as standby for the  master logger service. 

Master process manage the OS metrics data which is in CHM repository and manage the replica of the metrics information on the standby.


8) OSYSMOND : This process is responsible for monitoring and operating system metrics collection that sends data to cluster logger service  

Service run on all node in the cluster. This process send the  all metric data to the cluster logger service ologgerd


9) GPNPD : This is a small XML file containing information such as 

I) Cluster name

II) Location of VOTE disk

III) Network Information

IV) SPfile information for ASM

This file is individual for all nodes .If the node is already part of cluster agent reads the existing profile on the node. If node is being added to the cluster GPNP agent locates agent on the other existing nodes using multicast protocol and gets the profile from other agent nodes 


Please Follow the link To Install 2 node rac 19c


Hope This Helps ,

Thanks Sultan Khan

Previous
Next Post »

6 comments

Click here for comments
Unknown
admin
January 5, 2021 at 4:10 PM ×

Nice post easy to learn doc

Reply
avatar
Sid
admin
January 12, 2021 at 3:33 PM ×

East to understand

Reply
avatar
Unknown
admin
January 17, 2021 at 9:48 PM ×

Good post made it easy to understand

Reply
avatar