Oracle Grid Infrastructure - reconfiguration

There are some situations we need to reconfigure the Grid Home like IP-address change/hostname changer,
fresh GI installation due to binary issue etc..
Please follow the below steps


0. Please capture the below details before you do reconfigurations
      crsctl stat res -p
      crsctl stat res -t
      cluster Name:cemutlo -n
      scan name :  srvctl confif scan
      listener port :  srvctl confif listener
      host vip names
      if you are doing it in exadata   . please do the step 3 before you do the reconfiguration


1. Run the below command to deconfig the GI from all the nodes except the last node.

 $ORACLE_HOME/crs/install/rootcrs.pl -verbose -deconfig -force

2. Run the below command in the last node ( keepdg to retain the DG configuration)
 $ORACLE_HOME\crs\install\rootcrs.pl -deconfig -force -verbose -lastnode -keepdg





3. During reconfiguration, incase if the OCR disks are not listed as candidates for OCR DG creations, we may need to delete and add the disks

    ORACLE ASM:
            we can use either Plan A or Plan B         
          
           Plan A:
                   oracleasm createdisk <label> <dev_mapper>
                   oracleasm createdisk OCR_DK1  /dev/mapper/ASM_OCR_0001
                  
          Plan B:
                 incase the above is faling with "Unable to open device "/dev/mapper/ASM_OCR_0001": Device or resource busy "
          /usr/sbin/asmtool -C -l /dev/oracleasm -n OCR_DK1 -s  /dev/mapper/ASM_OCR_0001 -a force=yes

          Post task:
                   oracleasm scandisks
                   oracleasm listdisks
     AFD:
            ( BS = block size) and count= 1000 * BS = size of the disk
     
              dd if=/dev/zero of=/dev/mapper/ASM_REG_OCR_0001 bs=1M count=1000
        Post task:
       asmcmd afd_scan
       asmcmd afd_lsdsk
    Exadata:
             delete 3 disks from a DG from different cell to make them as a candidate for the OCR DG ( this step need to be done before the deconfigruation run)


4. Cleanup the gpnp files from all the nodes
  find <GRID_HOME>/gpnp/* -type f -exec rm -rf {} \;




5. run config.sh
 cd $ORACLE_HOME/crs/config
 ./config.sh

 feed the below values
    => cluster_name, scan name, port & Disable GNS if it is not used.
    => ensure all the host and VIPs are fetched automatically, if not, please add them manually.
    => Storage for OCR: create new DG and choose the disks candidates.
   

ACFS on Exadata

Oracle Automatic Storage Management Cluster File System (ACFS) on Exadata Database Machine:

Starting with Oracle Grid Infrastructure version 12.1.0.2, Oracle ACFS supports all database files and general purpose files on Oracle Exadata Database Machine running Oracle Linux on database servers.

The following database versions are supported by Oracle ACFS on Exadata Database Machine:

• Oracle Database 10g Rel. 2 (10.2.0.4 and 10.2.0.5)
• Oracle Database 11g (11.2.0.4 and higher)
• Oracle Database 12c (12.1.0.1 and higher)

The below steps are duly tested in Exadata X7 machines

Pre-requisite for ACFS creation

1. Verify if ACFS/ADVM module is loaded ( oracleacfs & oracleadvm )

dcli -g /home/oracle/dbs_group -l oracle 'lsmod | grep oracle'

 

Step by step instruction to create ACFS on Exadata

 

1. As ORAGRID : Create ASM volume to be used for ACFS file systems on FLASH01 DG
    

asmcmd volcreate -G FLASH01 -s 10G acfsvol

 

 

        2) As ORAGRID:  Capture the volume device name

asmcmd volinfo -G FLASH01 acfsvol

 

For example this case:

 

Diskgroup Name: FLASH01

 

         Volume Name: ACFSVOL

         Volume Device: /dev/asm/acfsvol-118

         State: ENABLED

         Size (MB): 10240

         Resize Unit (MB): 512

         Redundancy: MIRROR

         Stripe Columns: 8

         Stripe Width (K): 1024

         Usage:

         Mountpath: 

 

 

        3) As ORAGRID:  verify if the volume device name created on all the nodes

dcli -g /home/oracle/dbs_group -l oracle ‘s -l /dev/asm/acfsvol-118’

 

        4) As ORAGRID:  Verify if ADVM is enabled

crsctl stat res ora.FLASH01.ACFSVOL.advm -t

 

For example this case:

 

crsctl stat res ora.FLASH01.ACFSVOL.advm -t

--------------------------------------------------------------------------------

Name           Target  State        Server                   State details

--------------------------------------------------------------------------------

Local Resources

--------------------------------------------------------------------------------

ora.FLASH01.ACFSVOL.advm

               ONLINE  ONLINE       fraespou0101             STABLE

               ONLINE  ONLINE       fraespou0102             STABLE

               ONLINE  ONLINE       fraespou0103             STABLE

               ONLINE  ONLINE       fraespou0104             STABLE

--------------------------------------------------------------------------------

 

    

For example this case:

 

 

5) AS ORAGRID:  Create ACFS on ADVM  

/sbin/mkfs -t acfs /dev/asm/acfsvol-321 –b 4K

 

 

 

 6) As root - Create the mount points for the file systems and change ownership

 

dcli -g ~/dbs_group -l root ‘mkdir /u01/app/oracle/admin/common’   dcli -g ~/dbs_group -l root 'chown oracle:oinstall /u01/app/oracle/admin/common’   dcli -g ~/dbs_group -l root 'chmod 775 /u01/app/oracle/admin/common’

 

7) As ROOT:  Add filesystem  

/u01/app/oragrid/product/18.0/bin/srvctl add filesystem -d /dev/asm/acfsvol-321 -m /u01/app/oracle/admin/common  -u oracle -fstype ACFS -autostart ALWAYS

 

 

8) As ROOT/Oracle:  Start  filesystem  

/u01/app/oragrid/product/18.0/bin/srvctl start filesystem -d /dev/asm/acfsvol-321

 

 

 

  

Verifications

 

 

9) root/oragrid:  Verify if change ownership and permission  

crsctl stat res ora.FLASH01.ACFSVOL.acfs -t

 

crsctl stat res ora.FLASH01.ACFSVOL.advm –t

 

 

 

 

10) as Oracle:  list the ACFS FS  from all the nodes  

dcli -g  /home/oracle/dbs_group -l oracle df –h /u01/app/oracle/admin/common

 

 

 

 

 

11) as Oracle:  Touch a file on the ACFS FS  from all the nodes  

dcli -g  /home/oracle/dbs_group -l oracle touch /u01/app/oracle/admin/common/a.txt dcli -g  /home/oracle/dbs_group -l oracle cat /u01/app/oracle/admin/common/a.txt

 

 

 

How to restore a corrupted/missing datafile in standby:

How to restore a corrupted/missing datafile in standby:


Before 12.2
1. Cancel the recovery
    dgmgrl > edit database <db> set state='APPLY-OFF';
     (or) alter database recover managed standby database cancel;

2. From Primary take a datafile copy
rman target /
 backup as copy datafile 2 format "/u01/app/oracle/data01/sysaux_01.dbk" ;

3. scp the file from primay server to standby server ( may be to the same location in standby server)
4. catalog the datafile copy ( in sby server)
rman target /
  catalog datafilecopy '/u01/app/oracle/data01/sysaux_01.dbk'

5. Either do a switch copy or restore the datafile to the same location mentioned in the controlfile( in sby server)
   rman target /
 switch datafile 2 to copy;
 report schema

  alternatively
RUN {
  ALLOCATE CHANNEL ch00 TYPE disk ;
  restore datafile 2  ;
  }

6. enable recovery
    dgmgrl > edit database <db> set state='APPLY-ON';
     (or) alter database recover managed standby database disconnect from session ;


After 12
rman target sys/oracle@prod
RMAN> run
{
set newname for datafile 4 to '/u01/oracle/data02/users01.dbf' section size 1G;
restore (datafile 4 from service prodservice) using compressed backupset;
catalog datafilecopy '/u01/oracle/data02/users01.dbf';
}

OPatch lsinventory command doesnot show RAC details

Please use the below command to get the remote nodes patching details


 $ORACLE_HOME/OPatch/opatchauto report -format xml -type patches -remote


===============================================================
This is applicable to all releases of OPatch.
 The 12.2.0.1.13 release (and later) of OPatch, which is used for all releases 12.1.0.x and later
 The 11.2.0.3.18 release (and later) of OPatch, which is used for all releases 11.2.0.1 - 11.2.0.4
 Beginning with these releases of OPatch,  OPatch will only support patching/listing inventory for local node  of a RAC cluster. There will be no propagation to other nodes in cluster.


1. In Opatch 12.2.0.1.13 and Opatch 11.2.0.3.18, OPatch command option “-all_nodes” will be no-op and existed in -help.
     If opatch command is being called with option “-all_nodes”, Opatch will print out the warning on console msg as:
     ""OPatch was called with -all_nodes option. The -all_nodes option is being deprecated. Please remove it while calling OPatch."
 2.  In future release Opatch 12.2.0.1.14 (and later) and Opatch 11.2.0.3.20 (and later) , Option “-all_nodes” will be removed from -help, user will get syntax error if specify the option.
Alternative Feature of -all_node is available  for Multi-Node GI/RAC :

 $ORACLE_HOME/OPatch/opatchauto report -format xml -type patches -remote
(The remote command will get a list of patches from all nodes)
===========================================================


Refer:
GI/RAC/Single Instance Announcing Deprecation of OPatch Command Option "-all_nodes" (Doc ID 2331762.1)