This HowTo demonstrates migrating a RAC cluster to Flex ASM from 'traditional' ASM.
Flex ASM is, to a certain extent, to Database 12c what SCAN is to Database 11gR2. However, Flex ASM is more revolutionary than SCAN, which itself is more evolutionary. Even with non-clustered database installations you were always able to run more than one listener and use an address list to provide access while at least one listener was running. SCAN just allowed you to minimise the amount of listeners required on a large RAC cluster. Flex ASM, however, provides us with something that was previously unattainable: the ability to run multiple, independent in cardinality, ASM instances. This is just one of many different documented configurations.
The prerequisites are that the OCR files, ASM parameter file and ASM password file are stored in an ASM disk group with asm.compatible set to at least 12.1. We will confirm the ASM cluster mode before we begin and after we have finished.
[oracle@h1-o12-cl1 ~]$ . oraenv ORACLE_SID = [oracle] ? +ASM1 The Oracle base remains unchanged with value /u01/app/oracle [oracle@h1-o12-cl1 ~]$ asmcmd showclustermode ASM cluster : Flex mode disabled
Now run The ASM Configuration Assistant, asmca, and select the ASM Instances tab. You will be presented with the following table and a button to 'Convert to Oracle Flex ASM'.
| Node | Instance Name | Status | ADVM Driver Status |
|---|---|---|---|
| h1-o12-cl1 | +ASM1 | Up | Loaded |
| h2-o12-cl1 | +ASM2 | Up | Loaded |
Click the button and you will be presented with a table showing the network interfaces and their IPv4 addresses (IPv6 addresses are not shown).
| Check | Interface Name | Subnet | Interface In Use |
|---|---|---|---|
| No | eth0 | 10.1.2.0 | No-public |
| No | eth1 | 192.168.0.0 | Yes-cluster_interconnect |
Note that Flex ASM is limited to using an IPv4 network, so either create a new IPv4 network or check the 'eth1' network interface and click 'OK'. When ready, choose 'Yes' when informed that GI will be restarted one node at a time. You will then be given the name of a script to run on the local node, i.e. /u01/app/oracle/cfgtoollogs/asmca/scripts/converttoFlexASM.sh, containing the following commands.
#!/bin/sh
su oracle -c "/u01/app/12.1.0/grid/bin/srvctl add asm -proxy"
/u01/app/12.1.0/grid/srvm/admin/srvmhelper -updateDepASM
for node in h1-o12-cl1 h2-o12-cl1
do
/u01/app/12.1.0/grid/bin/crsctl stop cluster -n $node
/u01/app/12.1.0/grid/bin/crsctl start cluster -n $node
stat=1
cmd="/u01/app/12.1.0/grid/bin/crsctl status server $node"
while [ $stat = 1 ]
do
sleep 5
op=`$cmd`
op=`echo $op | /bin/cut -d' ' -f2 | /bin/cut -d'=' -f2`
if [ "$op" = "ONLINE" ]
then
stat=0
echo "Cluster bounced in node $node"
fi
done
done
Run this script as root.
[root@h1-o12-cl1 ~]# time /u01/app/oracle/cfgtoollogs/asmca/scripts/converttoFlexASM.sh CRS-2673: Attempting to stop 'ora.crsd' on 'h1-o12-cl1' CRS-2790: Starting shutdown of Cluster Ready Services-managed resources on 'h1-o12-cl1' [...] CRS-2672: Attempting to start 'ora.crsd' on 'h1-o12-cl1' CRS-2676: Start of 'ora.crsd' on 'h1-o12-cl1' succeeded Cluster bounced in node h1-o12-cl1 CRS-2673: Attempting to stop 'ora.crsd' on 'h2-o12-cl1' CRS-2790: Starting shutdown of Cluster Ready Services-managed resources on 'h2-o12-cl1' [...] CRS-2672: Attempting to start 'ora.crsd' on 'h2-o12-cl1' CRS-2676: Start of 'ora.crsd' on 'h2-o12-cl1' succeeded Cluster bounced in node h2-o12-cl1 real 4m42.838s user 0m1.658s sys 0m1.518s
Now recheck the cluster mode of ASM.
[oracle@h1-o12-cl1 ~]$ . oraenv ORACLE_SID = [oracle] ? +ASM1 The Oracle base remains unchanged with value /u01/app/oracle [oracle@h1-o12-cl1 ~]$ asmcmd showclustermode ASM cluster : Flex mode enabled
Note the new +APX proxy instance running now and the new listener endpoint on the private interconnect's IPv4 address.
[root@h2-o12-cl1 ~]# oifcfg getif eth0 10.1.2.0 global public eth1 192.168.0.0 global cluster_interconnect,asm eth0 2001:123:4:5:: global public
[oracle@h1-o12-cl1 ~]$ ps -ef|grep pmon_+
oracle 9098 1 0 17:50 ? 00:00:00 asm_pmon_+ASM1
oracle 10122 1 0 17:51 ? 00:00:00 apx_pmon_+APX1
[oracle@h2-o12-cl1 ~]$ ps -ef|grep tns
root 10 2 0 17:09 ? 00:00:00 [netns]
oracle 11821 1 0 17:53 ? 00:00:00 /u01/app/12.1.0/grid/bin/tnslsnr ASMNET1LSNR_ASM -no_crs_notify -inherit
oracle 12029 1 0 17:53 ? 00:00:00 /u01/app/12.1.0/grid/bin/tnslsnr LISTENER -no_crs_notify -inherit
oracle 12032 1 0 17:53 ? 00:00:00 /u01/app/12.1.0/grid/bin/tnslsnr LISTENER_SCAN1 -no_crs_notify -inherit
[oracle@h2-o12-cl1 ~]$ netstat -ntl|grep 1521
tcp 0 0 10.1.2.95:1521 0.0.0.0:* LISTEN
tcp 0 0 10.1.2.93:1521 0.0.0.0:* LISTEN
tcp 0 0 10.1.2.90:1521 0.0.0.0:* LISTEN
tcp 0 0 192.168.0.90:1521 0.0.0.0:* LISTEN
tcp 0 0 2001:123:4:5::3:60:1521 :::* LISTEN
tcp 0 0 2001:123:4:5::3:5d:1521 :::* LISTEN
tcp 0 0 2001:123:4:5:216:3ef:1521 :::* LISTEN
[oracle@h1-o12-cl1 ~]$ srvctl status asm -detail
ASM is running on h1-o12-cl1,h2-o12-cl1
ASM is enabled.
[oracle@h1-o12-cl1 ~]$ srvctl config asm
ASM home: /u01/app/12.1.0/grid
Password file: +CL1OCR/orapwASM
ASM listener: LISTENER
ASM instance count: ALL
Cluster ASM listener: ASMNET1LSNR_ASM
[oracle@h1-o12-cl1 ~]$ crsctl status resource -t
--------------------------------------------------------------------------------
Name Target State Server State details
--------------------------------------------------------------------------------
Local Resources
--------------------------------------------------------------------------------
ora.ASMNET1LSNR_ASM.lsnr
ONLINE ONLINE h1-o12-cl1 STABLE
ONLINE ONLINE h2-o12-cl1 STABLE
ora.proxy_advm
ONLINE ONLINE h1-o12-cl1 STABLE
ONLINE ONLINE h2-o12-cl1 STABLE
ora.asm
1 ONLINE ONLINE h1-o12-cl1 STABLE
2 ONLINE ONLINE h2-o12-cl1 STABLE
Let's now kill ASM on cluster node 1 and see what happens to the RAC database instance on node 1.
[oracle@h1-o12-cl1 ~]$ ps -ef|grep pmon oracle 9098 1 0 17:50 ? 00:00:00 asm_pmon_+ASM1 oracle 9423 1 0 17:50 ? 00:00:00 ora_pmon_cl1cdb1_1 oracle 10122 1 0 17:51 ? 00:00:00 apx_pmon_+APX1 oracle 16897 15751 0 18:42 ttyS0 00:00:00 grep pmon [oracle@h1-o12-cl1 ~]$ kill -9 9098 10122 [oracle@h1-o12-cl1 ~]$ crsctl stat res ora.asm -v NAME=ora.asm TYPE=ora.asm.type LAST_SERVER=h1-o12-cl1 STATE=OFFLINE TARGET=ONLINE CARDINALITY_ID=1 CREATION_SEED=92 RESTART_COUNT=1 FAILURE_COUNT=0 FAILURE_HISTORY= ID=ora.asm 1 1 INCARNATION=1 LAST_RESTART=NEVER LAST_STATE_CHANGE=08/26/2013 18:42:56 STATE_DETAILS= INTERNAL_STATE=CLEANING INSTANCE_COUNT=2 LAST_SERVER=h2-o12-cl1 STATE=ONLINE on h2-o12-cl1 TARGET=ONLINE CARDINALITY_ID=2 CREATION_SEED=92 RESTART_COUNT=0 FAILURE_COUNT=0 FAILURE_HISTORY= ID=ora.asm 2 1 INCARNATION=0 LAST_RESTART=NEVER LAST_STATE_CHANGE=NEVER STATE_DETAILS= INTERNAL_STATE=STABLE INSTANCE_COUNT=2
ASM has restarted and the database instance's PMON has remained stable.
[oracle@h1-o12-cl1 ~]$ ps -ef|grep pmon oracle 9423 1 0 17:50 ? 00:00:00 ora_pmon_cl1cdb1_1 oracle 16956 1 0 18:43 ? 00:00:00 asm_pmon_+ASM1 oracle 17130 1 0 18:43 ? 00:00:00 apx_pmon_+APX1
The alert log for this instance shows what transpired during this reconnection process.
Mon Aug 26 18:43:13 2013 NOTE: ASMB registering with ASM instance as client 0x10000 (reg:95210492) NOTE: ASMB connected to ASM instance +ASM2 (Flex mode; client id 0x10000) NOTE: ASMB rebuilding ASM server state NOTE: ASMB rebuilt 2 (of 2) groups NOTE: ASMB rebuilt 19 (of 19) allocated files NOTE: fetching new locked extents from server NOTE: 0 locks established; 0 pending writes sent to server SUCCESS: ASMB reconnected & completed ASM server state Mon Aug 26 18:44:25 2013 ALTER SYSTEM SET local_listener=' (ADDRESS=(PROTOCOL=TCP)(HOST=2001:123:4:5::33:5c)(PORT=1521))','(ADDRESS=(PROTOCOL=TCP)(HOST=10.1.2.92) (PORT=1521))' SCOPE=MEMORY SID='cl1cdb1_1';
But a short while later, CRS performs a normal shutdown of this database instance...