Why and How to install Grid 19c on RHEL 8?

 Why ?

Simply we will be requested to install Oracle Grid RAC DB on Redhat RHEL 8, below is my note for this installation.

How ?

1-  OS installed + asm disks by servers admin

Done By OS and VM team
Ensure: secure boot not working for grid

# mokutil --sb-state

2- Install packages:

https://docs.oracle.com/en/database/oracle/oracle-database/19/cwlin/supported-oracle-linux-8-distributions-for-x86-64.html

3- setup ASMLIB:

## linux disks for redhat :

https://access.redhat.com/solutions/315643

https://labmice.in/oraclerac/RHEL8.10/Prerequsites_OS_Oracle_RAC_RHEL8.10/

OL8/RHEL8: ASMLib: root.sh is failing with CRS-1705: Found 0 Configured Voting Files But 1 Voting Files Are Required (Doc ID 2789052.1)

 

yum install kmod-oracleasm

yum localinstall libbpf-0.6.0-6.el8.x86_64.rpm

yum localinstall oracleasmlib-2.0.12-1.el7.x86_64.rpm

yum localinstall oracleasm-support-2.1.11-2.el7.x86_64.rpm

 

https://docs.oracle.com/en/operating-systems/oracle-linux/asmlib/asmlib-ConfiguringASMLib.html#querying_asm_disk_information

 

lsblk

parted /dev/sde --script mklabel gpt mkpart primary 0% 100%

sudo oracleasm configure -i

sudo systemctl enable --now oracleasm

sudo oracleasm createdisk data1 /dev/sde1

sudo oracleasm scandisks

oracleasm listdisks

4- setup DNS records

**3 IPs for scan

** 2 IPs for vir ( added to hosts )

** hosts in DNS

5- creat directories and user for Grid

create user/groups

  groupadd -g 54321 oinstall

  groupadd -g 54329 asmadmin

  groupadd -g 54327 asmdba

  groupadd -g 54328 asmoper

  groupadd -g 54322 dba

  groupadd -g 54323 oper

  groupadd -g 54324 backupdba

  groupadd -g 54325 dgdba

  groupadd -g 54326 kmdba

  groupadd -g 54330 racdba

useradd -u 54321 -g oinstall -G dba,asmdba,backupdba,dgdba,kmdba,racdba,racdba grid

useradd -u 54331 -g oinstall -G dba,asmdba,backupdba,dgdba,kmdba,racdba,asmadmin,asmoper,oper oracle

echo "oracle123" | passwd --stdin oracle

echo "grid123" | passwd --stdin grid

 

## create directories

mkdir -p /u01/app/19.0.0/grid

mkdir -p /u01/app/grid

chown -R grid:oinstall /u01

chmod -R 775 /u01/

 

## etc hosts on both nodes

#Virtual IPs:

x.x.x.x  <hostname1>-vip   <hostname1>-vip.<domain>

x.x.x.x  <hostname2>-vip   <hostname2>-vip.<domain>

6- unzip home and apply patch during the instllation:

## patch GRID_HOME with latest page during the installation ** make sure to update OPatch version

./gridSetup.sh -applyRU /u01/patches/34762026

/u01/app/19.0.0/grid/gridSetup.sh -applyRU /u01/files/34762026 (GI Patch)

 

Why and How to know and compress Table as Hybrid Columnar Compression (HCC)?

Why ?

To get the most compress advance feature on the Oracle Exadata machine you may use the Hybrid Columnar Compression(HCC), however , there is a limitation on it where you have to use it for a certain object type, moreover, the way to apply this is online and offline , below I will simply mention the query to identified these objects and the option of doing this.

How ?

Below is a query that will identify and create executable scripts for the offline way where table will be locked, and indices will be unusable after doing this.

1- Identify and generate offline script for HCC:

select  -

'ALTER TABLE '||table_owner||'.'||table_name||' move PARTITION '||PARTITION_NAME||' compress for query high ;' Result

from dba_TAB_PARTITIONS

where (COMPRESS_FOR<> 'QUERY HIGH' or COMPRESS_FOR is null)

and table_name not in (select table_name from dba_tab_columns

where data_type in ('CLOB','LOB','LONG', 'BLOB', 'LONG RAW'))

and table_name not in(select TABLE_NAME from dba_indexes

where index_type in ('BITMAP')

)and TABLE_NAME not in(select OBJECT_NAME from dba_recyclebin dr where dr.OWNER=TABLE_OWNER)

--and TABLE_OWNER NOT IN ('SYS','SYSTEM')

ORDER BY TABLE_OWNER,TABLE_NAME,PARTITION_NAME ;

2- Using DBMS_REDEFINITION.REDEF_TABLE for the online HCC implementation:

BEGIN

  DBMS_REDEFINITION.REDEF_TABLE (

    uName                      => '<table_owner>'

   ,tName                      => '<table_name>'

   ,table_compression_type     => 'COLUMN STORE COMPRESS FOR QUERY HIGH');

END;

/

 

Create a new Server DB using Existing ASM disks - DB Disk replication ASM

Why?

In an assumption that computing DB server got crashed and only ASM disks which are in storage appliance are available or in a case of data replication using ASM disk storage replication which not a recommended approach but I am here just elaborate how you will achieve this in case of needed.

How?

In my case the original DB is RAC two nodes and I will create a new single node to using that ASM disks and start the DB using these disks in a new computing node with version 19c.

1- Install both Grid and DB as software only using same version and patch level of the orginal one:

2- Start using "root" the HAS for oracle restart:

. oraenv

/u01/app/19.0.0/grid/perl/bin/perl -I/u01/app/19.0.0/grid/perl/lib/ -I/u01/app/19.0.0/grid/crs/install/ /u01/app/19.0.0/grid/crs/install/roothas.pl

 

3- using grid start the other service as below.

crsctl start res ora.cssd -init

 

4- From old  nodes:

 get the asm parameter file and change what is needs to be changed,

get oracle parameter file and change what is needs to be changed (cluster=no):

check the crs status and check diskgroups :

crsctl stat res –t

ASMCMD> lsdg

start ASM whether using sqlplus / as sysasm

or 

using after add the resource:

srvctl start asm

 

5- start DB using pfile:

#sqlplus / as sysdba

SQL> startup nomount pfile='initorcla.ora' ;

SQL> alter database mount;

SQL> alter database open;

 

6- add needed resources in the SRVCTL to be part of Grid:

srvctl add listener -l LISTENER -p "TCP:1541" -o $ORACLE_HOME

srvctl add asm -l LISTENER -p $ORACLE_HOME/dbs/init+ASM.ora

srvctl add database -d <dbname> -o $ORACLE_HOME -p /export/home/oracle/pfile_<dbname>.ora -pwfile +DATA/ICM/PASSWORD/pwd<dbname>.257.1132500007

crsctl stat res -t

 

below references may help with troubleshooting and other details :

https://eleoracle.wordpress.com/2015/01/23/move-asm-diskgroups-between-server/

https://www.br8dba.com/cssd-wont-start-automatically/

Rename the Database

Why?

One of the common change for database for the Data Guard implementation is to match the DB name and have a different DB unique name, it is article I will list my own steps to keep achieve this along with other changes to avoid any issue post to this change. In my below steps I have match the oracle name and oracle SID to follow the same pattern of default installation values.

How?

With reference to the blog https://oracle-base.com/articles/9i/dbnewid , I did my below steps.

1- Check the current value for the SID and database name:

SQL> show parameter name ;

SQL> select instance from v$thread;

SQL> select name from v$database;

 

2- stop the DB and started on mount stat:

SQL> SHUTDOWN IMMEDIATE

SQL> STARTUP MOUNT

 

3- rename the DB using “nid” utility, setname parameter to only change the name and keep the DB id, shutdown it and started on mount to change parameter file , error will appeared but ignore it.

$ nid TARGET=sys/sys@<current_service_name> DBNAME='<new_name>' SETNAME=YES

SQL> SHUTDOWN IMMEDIATE

SQL> STARTUP MOUNT

SQL> ALTER SYSTEM SET DB_NAME='<new_name>' SCOPE=SPFILE;

SQL> SHUTDOWN IMMEDIATE

 

4- change the password file and spfile name to match the new name:

$ orapwd file=orapw$ORACLE_SID password=sys force=y

Export ORACLE_SID=<new_DBNAME>;

SQL> STARTUP MOUNT

--*Rename the SPFILE to match the new DBNAME

$ cd $ORACLE_HOME/dbs

$ cp spfile<old_dbname>.ora spfile$ORACLE_SID.ora

$ lsnrctl reload

SQL> STARTUP MOUNT ;

SQL> show parameter name ;

SQL> alter database open ;

 

Why and How to install RAC DB 19c on Solaris 11.4 ?

 Why ?

RAC DB is the most powerful RDBMS DB for high performance and high availability. It is an expected task for any Oracle DBA to do RAC installation and here I am sharing my quick reference while do this task along with some details related to disks and OS configuration to achieve this.

How ?

 

Below are steps that I follow to do this installation where I have the below:

Download GI+DB binary for Solaris Sparc64

Download the latest GI patch that will include both GI and DB and make sure you have opatch latest one as well :

1- OS pre configuration:

## check NTP

svcs -xv ntp

svcs /network/ntp:default

svcs -l /network/ntp:default

ntpq -p

https://docs.oracle.com/cd/E26502_01/html/E28996/time-20.html

 

#TO setup the slew

svccfg -s svc:/network/ntp:default setprop config/slew_always = true

svcadm refresh svc:/network/ntp:default

#To check it :

svcprop -p config/slew_always svc:/network/ntp:default

svccfg -s svc:/network/ntp:default listprop | /bin/grep slew

## check DNS

nslookup

## install packages

root@node1:~# pkg list entire

NAME (PUBLISHER)                                  VERSION                    IFO

entire                                            11.4-11.4.52.0.1.132.2     i--

root@node1:~# pkg install oracle-database-preinstall-19c

2- Users/group setup:

## create user/groups

  groupadd -g 54321 oinstall

  groupadd -g 54329 asmadmin

  groupadd -g 54327 asmdba

  groupadd -g 54328 asmoper

  groupadd -g 54322 dba

  groupadd -g 54323 oper

  groupadd -g 54324 backupdba

  groupadd -g 54325 dgdba

  groupadd -g 54326 kmdba

  groupadd -g 54330 racdba

 

useradd -u 54321 -g oinstall -G dba,asmdba,backupdba,dgdba,kmdba,racdba,racdba -d "/export/home/oracle" -m oracle

useradd -u 54331 -g oinstall -G dba,asmdba,backupdba,dgdba,kmdba,racdba,asmadmin,asmoper -d "/export/home/grid" -m grid

3-User project in OS:

# to check current configured project

cat /etc/project

 

## using project to create users:

projadd -p 100 -G dba -c "Oracle Project" -K "project.max-shm-memory=(privileged,10G,deny)" group.dba

projmod -sK "project.max-sem-ids=(privileged,100,deny)" group.dba

 projmod -sK "project.max-shm-ids=(privileged,100,deny)" group.dba

  projmod -sK "project.max-sem-nsems=(privileged,256,deny)" group.dba

projmod -sK"process.max-stack-size=(basic,10485760,deny)" group.dba

 

useradd -u 54321 -g oinstall -G dba,asmdba,backupdba,dgdba,kmdba,racdba,racdba -d "/export/home/oracle" -m -K project=group.dba oracle

useradd -u 54331 -g oinstall -G dba,asmdba,backupdba,dgdba,kmdba,racdba,asmadmin,asmoper -d "/export/home/grid" -m -K project=group.dba grid

 

## to create a project and then add the user use the below :

 

projadd -G dba -c "Oracle dba Group for CRS and RDBMS" group.dba

projmod -sK "project.max-shm-memory=(priv,32G,deny)" group.dba

projmod -sK "project.max-sem-ids=(priv,256,deny)" group.dba

projmod -sK "process.max-sem-nsems=(priv,256,deny)" group.dba

projmod -sK "process.max-file-descriptor=(priv,65536,deny)" group.dba

projmod -sK "process.max-file-descriptor=(basic,65536,deny)" group.dba

projmod -sK "process.max-stack-size=(basic,10485760,deny)" group.dba

usermod -K project=group.dba grid

usermod -K project=group.dba oracle

4- Create directories:

## create directories

mkdir -p /u01/app/19.0.0/grid

mkdir -p /u01/app/grid

chown -R grid:oinstall /u01

chmod -R 775 /u01/

 

5- Create directories:

## create directories

mkdir -p /u01/app/19.0.0/grid

mkdir -p /u01/app/grid

chown -R grid:oinstall /u01

chmod -R 775 /u01/

 

6- prepare ASM disks:

## Some references for the disk setup :

 

AFD Storage Disks Not Appearing Under "CREATE ASM DISK GROUP" In 19c Grid Installation (Doc ID 2878933.1)

Configuring Disk Devices for Oracle ASM on Oracle Solaris

https://support.oracle.com/epmos/faces/CommunityDisplay?resultUrl=https%3A%2F%2Fcommunity.oracle.com%2Fmosc%2Fdiscussion%2Fcomment%2F14658232&_afrLoop=72302341112795&resultTitle=Re%3A+ASMFD+%3A+before+installing+GRID+use+of+asmfd_label+seems+to+do+nothing.+AFD_STATE+showing+nothing.&commId=&displayIndex=1&_afrWindowMode=0&_adf.ctrl-state=wiy4fphbq_634#Comment_14658232

 

You have to get the list of shared disk from OS Admin include the cylinder to be used, below I will show samples only :

 

you may use command “format” to see all disks details, below is a sample of initiate and label the disks.

 

After unzip the GI home by following the below link, start the labeling.

https://docs.oracle.com/en/database/oracle/oracle-database/19/cwsol/installing-oracle-standalone-cluster.html#GUID-7BAE8D62-4D26-424C-BBB4-B993F248C163

 

chown grid:asmadmin /dev/rdsk/c0t60000970000297801140533030303344d0s0

chown -h grid:asmadmin /dev/rdsk/c0t60000970000297801140533030303344d0s0

chmod g+w /devices/scsi_vhci/disk@g60000970000297801140533030303344:a,raw

ls -ltrh /dev/rdsk/c0t60000970000297801140533030303344d0s0

ls -ltrh /devices/scsi_vhci/disk@g60000970000297801140533030303344:a,raw

 

DATA Disk Group

As root user :

export ORACLE_HOME=/u01/app/19.0.0/grid

export ORACLE_BASE=/u01/app/grid

cd /u01/app/19.0.0/grid/bin

 

./asmcmd afd_label MGMT1 /dev/rdsk/c0t60000970000297801140533030303344d0s0 –init

./asmcmd afd_lslbl /dev/rdsk/c0t60000970000297801140533030303344d0s0

 

https://support.oracle.com/epmos/faces/DocumentDisplay?_afrLoop=368998006909019&parent=EXTERNAL_SEARCH&sourceId=PROBLEM&id=2355039.1&_afrWindowMode=0&_adf.ctrl-state=168weusfr1_4

 

 

In case you do a deinstall, below will help you to wipe and clean up the disks :

 

dd if=/dev/zero of=/dev/rdsk/c0t60000970000297801140533030303344d0s0 bs=1024 count=100

./asmcmd afd_unlabel /dev/rdsk/c0t60000970000297801140533030303344d0s0 –init

 

 

7- start GI instllation:

## To avoid any issue and bugs ,I do prefer to patch the GI during the installation with the latest RU .

## to avoid GI installation issue , chwon to grid after setup the disks using root ,some logs will be owned by root and this will affect the GI installation:

 chown -R grid:oinstall /u01/app/19.0.0/grid/log

 

You may run cluvy as well prior to installation or you may depned on GUI cluvy during the installation.

## Cluvy command , you have to install clufy separately

cluvfy stage -pre crsinst -n node1,node2 -fixup -verbose

 

 

## patch GRID_HOME with latest page during the installation ** make sure to update OPatch version

 

## run installer from Grid user home

/u01/app/19.0.0/grid/gridSetup.sh -applyRU /u01/files/34762026

 

Follow the wizard till the scripts step by step.

 

8- DB instllation:

mkdir -p /u02/app/oracle

mkdir -p /u02/app/oracle/product/19.0.0/dbhome_1

cd /u02/app/oracle/product/19.0.0/dbhome_1

unzip -q /export/home/sal3570/oracle_files/DB_19c_instllation.zip

#update opatch and prepare DB, OCW patches to be applied during installation

./runInstaller -applyRU /u02/patches/34762026/34765931 -applyOneOffs /u02/patches/34762026/34768559

 

Follow the wizard till the scripts step by step.

 

9- GI wizard sample :

 

10- DB wizard sample :

## Then DBCA 

## this should be confirmed with the Application requirement