3.7.3 Configuring Oracle Exadata Storage Servers After Rescue

After a successful rescue, you must configure the cell. If the data partitions were preserved, then the cell disks were imported automatically during the rescue procedure.

1. Re-create the cell disks and grid disks for any disks that were replaced during the rescue procedure.
   1. Create the cell disks on only the replaced disks using the following command:

      # cellcli -e create celldisk all harddisk

   2. Get the name of the new cell disks created.
   3. Get the mapping of the grid disks.

      cellcli -e list griddisk attributes name,offset,size

      Get the grid disk attributes from an existing disk. If you replaced a system disk (slot 0 or slot 1 on X6 and earlier servers), then you need to retrieve the values from the other system disk. If any of the grid disks are SPARSE grid disks, then also get the virtualsize attribute from another sparse disk.

      For example, if the new grid disks are CD_01* and CD_08*, then you would use a command such as this:

      # cellcli -e list griddisk attributes name,cachingpolicy,offset,size,virtualsize |
      egrep '_CD_00|_CD_07'
      DATAC1_CD_00_dbm01celadm04    default    32M                  779G
      DATAC1_CD_07_dbm01celadm04    default    32M                  779G
      DBFSC1_CD_07_dbm01celadm04    default    1.0575714111328125T  33.6875G
      RECOC1_CD_00_dbm01celadm04    none       887.046875G          195.90625G
      RECOC1_CD_07_dbm01celadm04    none       887.046875G          195.90625G
      SPARSEC1_CD_00_dbm01celadm04  default    779.046875G          108G         1.0546875T
      SPARSEC1_CD_07_dbm01celadm04  default    779.046875G          108G         1.0546875T

   4. Create the grid disks on the new cell disks using the retrieved attributes.

      For example, using the attributes retrieved in the previous step for CD00, you would create the grid disks on CD01 as follows:

      # cellcli -e create griddisk DATAC1_CD_01_dbm01celadm04 celldisk=CD_01_dbm01celadm04, 
      size=779G, cachingpolicy=default
      
      # cellcli -e create griddisk SPARSEC1_CD_01_dbm01celadm04 celldisk=CD_01_dbm01celadm04, 
      size=108G, virtualsize=1.0546875T,cachingpolicy=default
      
      # cellcli -e create griddisk RECOC1_CD_01_dbm01celadm04 celldisk=CD_01_dbm01celadm04 , 
      size=195.90625G, cachingpolicy=none

      Using the attributes retrieved in the previous step for CD07, you would create the grid disks on CD08 as follows:

      # cellcli -e create griddisk DATAC1_CD_08_dbm01celadm04 celldisk=CD_08_dbm01celadm04, 
      size=779G, cachingpolicy=default
      
      # cellcli -e create griddisk SPARSEC1_CD_08_dbm01celadm04 celldisk=CD_08_dbm01celadm04, 
      size=108G, virtualsize=1.0546875T,cachingpolicy=default
      
      # cellcli -e create griddisk RECOC1_CD_08_dbm01celadm04 celldisk=CD_08_dbm01celadm04, 
      size=195.90625G, cachingpolicy=none
      
      # cellcli -e create griddisk DBFSC1_CD_08_dbm01celadm04 celldisk=CD_08_dbm01celadm04, 
      size=33.6875G, cachingpolicy=default
2. Check the status of the grid disk.
   If any grid disk is inactive, alter its status to active.

   CellCLI> ALTER GRIDDISK ALL ACTIVE

3. Log in to the Oracle Automatic Storage Management (Oracle ASM) instance, and set the disks to ONLINE for each disk group:

   SQL> ALTER DISKGROUP disk_group_name ONLINE DISKS IN FAILGROUP \
   cell_name WAIT;

   Note:

   • If the command fails because the disks were already force-dropped, then you need to force-add the disks back to the Oracle ASM disk groups.

   • The grid disk attributes asmmodestatus and asmdeactivationoutcome will not report correctly until the ALTER DISKGROUP statement is complete.

4. Reconfigure the cell using the ALTER CELL command.

   In the following example, e-mail notification is configured to send notification messages to the storage server administrator according to the specified notification policy:

   CellCLI> ALTER CELL                                     -
              mailServer='mail_relay.example.com',            -
              smtpFromAddr='[e-mail]',         -
              smtpToAddr='[e-mail]',         -
              notificationPolicy='critical,warning,clear', -
              notificationMethod='mail,snmp'

5. Re-create the I/O Resource Management (IORM) plan.
6. Re-create the metric thresholds.