Table of Contents
Tested Version
- CRS – version : 11.2.0.4
Verify OCR backup and check OCR location
Locate OCR backups [root@grac42 ~]# ocrconfig -showbackup auto grac42 2014/10/01 12:53:10 /u01/app/11204/grid/cdata/grac4/backup00.ocr grac42 2014/10/01 08:53:09 /u01/app/11204/grid/cdata/grac4/backup01.ocr grac42 2014/09/30 18:41:00 /u01/app/11204/grid/cdata/grac4/backup02.ocr grac42 2014/09/30 09:16:23 /u01/app/11204/grid/cdata/grac4/day.ocr grac42 2014/09/16 09:34:38 /u01/app/11204/grid/cdata/grac4/week.ocr --> After an OCR corruption we need to restore: /u01/app/11204/grid/cdata/grac4/week.ocr Locate our OCR location [root@grac42 ~]# ocrcheck Status of Oracle Cluster Registry is as follows : Version : 3 Total space (kbytes) : 262120 Used space (kbytes) : 4040 Available space (kbytes) : 258080 ID : 630679368 Device/File Name : +OCR Check OCR DG content ASMCMD> ls -l OCR/grac4/asmparameterfile/ Type Redund Striped Time Sys Name ASMPARAMETERFILE MIRROR COARSE JUN 21 10:00:00 Y REGISTRY.253.850819057 N spfileASM.ora => +OCR/grac4/asmparameterfile/REGISTRY.253.850819057 ASMCMD> ls -l OCR/grac4/OCRFILE/ Type Redund Striped Time Sys Name OCRFILE MIRROR COARSE OCT 02 09:00:00 Y REGISTRY.255.828888017
Restore OCR from weekly backup
Stop the CRS stack on all nodes and restart ASM in exclusive mode on grac42 [root@grac42 ~]# crsctl stop crs [root@grac42 ~]# crsctl stop crs [root@grac42 ~]# $GRID_HOME/bin/crsctl start crs -excl -nocrs Remove the corrupted OCR file ASMCMD> cd +OCR/grac4/OCRFILE ASMCMD> rm REGISTRY.255.828888017 Restore OCR [root@grac42 ~]# $GRID_HOME/bin/ocrconfig -restore /u01/app/11204/grid/cdata/grac4/week.ocr ASMCMD> pwd +OCR/grac4/OCRFILE ASMCMD> ls -l Type Redund Striped Time Sys Name OCRFILE MIRROR COARSE OCT 02 10:00:00 Y REGISTRY.255.859888799 Restart CRS on all nodes any verify database startup [root@grac42 ~]# crsctl stop crs [root@grac42 ~]# crsctl stop crs [root@grac43 ~]# crsctl start crs [grid@grac42 bin]$ crs | egrep 'NAME|---|db' NAME TARGET STATE SERVER STATE_DETAILS ------------------------- ---------- ---------- ------------ ------------------ ora.grac4.db ONLINE ONLINE grac42 Open ora.grac4.db ONLINE ONLINE grac43 Open
Hi Helmut
In your great example you stop crs on the nodes. I can see you do it on the same node twice
[root@grac42 ~]# crsctl stop crs
[root@grac42 ~]# crsctl stop crs
[root@grac42 ~]# $GRID_HOME/bin/crsctl start crs -excl -nocrs
Should this be on grac42 and grac43?