RAC Attack - Oracle Cluster Database at Home/RAC Attack 12c/Convert to Flex ASM

Check the ASM flex status -

[grid@vm1 ~]$ asmcmd showclustermode ASM cluster : Flex mode disabled

Now you can fireup the ASMCA to convert this to a flex ASM

Flex ASM enables ASM instances to be run on separate physical server from the database servers. The DGs are managed by a small set of ASM instances running on a subset of cluster nodes. --- the cardinality

Flex ASM requires one or more designated network interfaces for communication among the ASM instances. Dedicated ASM Listener is configured for ASM instance communication on the ASM network.

You can have dedicated network for ASM communication or you can share the same network as the interconnect.

There is a mixed mode of flex ASM also - where the ASM instances are run on all the nodes. This option is used in case you need to user DB's with versions lower than 12.1

Conversion requires GRID INFRA to be restarted one node at a time. then it would ask to run the following script on the local node -

/u01/app/grid/cfgtoollogs/asmca/scripts/converttoFlexASM.sh After running the script the flex mode is enabled, however, clusterware would need to be restarted on both the nodes.

[grid@vm1 ~]$ asmcmd showclustermode Connected to an idle instance. ASM cluster : Flex mode enabled              <<<<<<<<<<<<<<  says so on both the nodes

However, as recommended, restart clusterware on both nodes.

hold that thought, it seems the script is doing that on its own.

Once the script completes, it's done -

[grid@vm1 ~]$ asmcmd showclustermode ASM cluster : Flex mode enabled

[grid@vm2 ~]$ asmcmd showclustermode ASM cluster : Flex mode enabled

So, that's it. No need to restart. It was restarted for you.

Following resource added to the cluster -

ora.ASMNET1LSNR_ASM.lsnr ONLINE ONLINE       vm1                      STABLE ONLINE ONLINE       vm2                      STABLE

[grid@vm2 ~]$ srvctl config asm ASM home:  Password file: +CRS/orapwASM ASM listener: LISTENER ASM instance count: 3                                <<<<<<<<<<<<<   this is the default cardinality Cluster ASM listener: ASMNET1LSNR_ASM

For now the following instances are running -

grid    22426     1  0 03:55 ? 00:00:00 asm_pmon_+ASM1    -- vm1 grid    17905     1  0 03:58 ? 00:00:00 asm_pmon_+ASM2     -- vm2

Yes, proxy instances are not running. Because they arent supposed to run. No ACFS

via command line run asmca in silent mode -

asmca -silent -convertToFlexASM -asmNetworks interface_name/subnet, interface_name/subnet... 			 -asmListenerPort ASM_listener_port_number

=
================== REDUCE CARDINALITY

=
==================

[grid@vm1 ~]$ srvctl modify asm -count 1                                         <<<<<<<<<<<  can reduce the cardinality. 2 is minimum PRCA-1123 : The specified ASM cardinality 1 is less than the minimum cardinality of 2. [grid@vm1 ~]$ srvctl modify asm -count 2

Earlier

ora.asm 1       ONLINE  ONLINE       vm1                      Started,STABLE 2       ONLINE  ONLINE       vm2                      Started,STABLE 3       OFFLINE OFFLINE                               STABLE                  <<<<<<<<< I'd only 2 nodes and cardinality was 3, no node for this to run

after reducing cardinality, the resource was removed from OCR ora.asm 1       ONLINE  ONLINE       vm1                      Started,STABLE 2       ONLINE  ONLINE       vm2                      Started,STABLE