Oracle Corner

Pre-checks

On each node

Labels: 11gR2 PSU5 Oracle RAC Patch DR server

// posted by Tatjana (Tanja) Kane @ 4/10/2012 0 comments

• Checks
• cemutlo -n; srvctl config scan (list cluster and scan name)
• crsctl check crs (summary check of top components)
• crsctl status resource -t (list detailed resources)
• crsctl query crs activeversion (check active version)
• srvctl status nodeapps/scan
• cluvfy comp crs  (verify CRS integrity)

• OCR
• crsctl query css votedisk 
• ocrconfig -replace ocrmirror (add a mirrored copy)
• ocrconfig -add +ASM_DG (add OCR to ASM disk group)
• ocrconfig -showbackup (OCR automatically backed up every 4 hours, at the end of a day, and at the end of a week). Check $CRS_HOME/cdata/$cluster_name
• ocrconfig -manualbackup

• Various other commands
• crsctl stop cluster -all (without "-all", it stops CS on the current node)
• crsctl disable crs (disables HA)
• crsctl stop crs (stops the cluster on the local node)
• srvctl stop instance -d DBNAME - i "INST1, INST2" -o immediate (also stops services)
• srvctl stop database -d DBNAME
• srvctl disable service ... 
• SQL>alter system kill session 'sid, serial, instance_id'; 

• ASM:
• srvctl status asm
• /etc/init.d/oracleasm listdisks or querydisks
• Add additional disks:
• oracleasm configure -i (to configure ASM library software)
• oraceasm scandissk (to make the disks available to other nodes)
• /etc/init.d/oracleasm createdisk DISK1 /dev/dasdxxxx (mark a disk as SAM disk)
• ASM config file in /etc/sysconfig/oracleasm

• Monitoring/diagnostics
• $Grid_home/bin/diagcollection.pl ... (collect diag information from CW)
• oclumon debug log deamonmodule all/osysmond/ologgerd/client debuglevel 0-3
• oclumon dumpnodeview (dump node view for all or some nodes, for time interval. This will list system usage, top consumer, processes, devices, amongst all)
• crsctl set trace "crs=1" (crs/cfs/css/evm, and level can be 0-5)
• export SRVM_TRACE=TRUE
• The column CLUSTER_WAIT_TIME in V$SQLAREA represents the wait time incurred by individual SQL for global chache events.
• Check AWR report and RAC related events:
• gc current block 2-way (block requested to be accessed in the current mode (U/D), sent by the master instance via Cache Fusion)
• gc current block busy (block being used by another instance, or holding instance couldn't write it fast enough)
• gc buffer busy acquire/release (block busy is now split into acquire and release)
• gc current/cr block congested (high load, CPU saturation)

// posted by Tatjana (Tanja) Kane @ 4/03/2012 0 comments

• runcluvfy.sh (post hw & sw, pre-installation)
• Check/set IP address resolution
• Set up SSH connectivity
• Create directories in /opt for the software
• Check ASMLib is configured
• Set .profile

• Install Grid Infrastructure and database software
• No shutdown necessary, it will happen automatically (rolling fashion?)
• Patch as root "opatch auto ... -ocmrf"
• Install the 11g OEM agent (chmod 750 directory)
• Patch the agent
• Create the DR database (easy with grid control)

Labels: Oracle RAC 11gR2 installation Linux DR

// posted by Tatjana (Tanja) Kane @ 4/03/2012 0 comments