This KB describes how to remove SVM devices in Oracle Solaris and also from diskset which is under control Oracle Solaris Cluster.
Step 1: Check Current Configurations
#scstat -g — Resource Groups — Group Name Node Name State Suspended ———- ——— —– ——— Group: xxxxbdb-rg xxxxbdbsunp00 Online Yes Group: xxxxbdb-rg xxxxbdbsunp01 Offline Yes Group: xxxfra-rg xxxxbdbsunp00 Online Yes Group: xxxfra-rg xxxxbdbsunp01 Offline Yes # metaset|grep -w “Set name” Set name = <disksetname>, Set number = 1 Set name = <disksetname>, Set number = 2 #metaset -s <disksetname> #metaset -s <disksetname>
-After you checked all configurations ,you should get an explorer for a failed operation.
#/opt/SUNWexplo/bin/explorer
Step 2: Find which SVM device ‘ll be removed
-If you need to remove a mount point which is under control of cluster resource , first you need to remove this mount point in cluster resource confiurations.
#clresource show -v|grep global FilesystemMountPoints: /global/X1 /global/X2 /global/X3 /global/X4[
-We ‘ll perform this operation for X3 and X4 filesystems.
-Check “df -h” output;
#df -h |egrep ‘X3|X4’ /dev/md/xxxphsds/dsk/d56 454G 38G 412G 9% /global/X3 /dev/md/xxxphsds/dsk/d155 236G 105G 129G 46% /global/X4
-We ‘ll remove SVM devices which is D56 and D155
# metastat -s <disksetname> d155 xxxphsds/d155: Mirror Submirror 0: xxphsds/d156 State: Okay Pass: 1 Read option: roundrobin (default) Write option: parallel (default) Size: 503162880 blocks (239 GB) xxxphsds/d156: Submirror of xxxphsds/d155 State: Okay Size: 503162880 blocks (239 GB) Stripe 0: Device Start Block Dbase State Reloc Hot Spare d79s0 0 No Okay No Device Relocation Information: Device Reloc Device ID d79 No – # metastat -s oraphsds d56 xxxphsds/d56: Mirror Submirror 0: xxxphsds/d57 State: Okay Pass: 1 Read option: roundrobin (default) Write option: parallel (default) Size: 967618560 blocks (461 GB) xxxphsds/d57: Submirror of xxxphsds/d56 State: Okay Size: 967618560 blocks (461 GB) Stripe 0: (interlace: 32 blocks) Device Start Block Dbase State Reloc Hot Spare d67s0 0 No Okay No d65s0 0 No Okay No Device Relocation Information: Device Reloc Device ID d67 No – d65 No – # metastat -s <disksetname> d56
Step 3: Remove SVM Device.
-First check your cluster state . It must be suspended and there musn’t be any process under the filesystems which you ‘ll perform unmount operations.
#scstat -g
#clrg suspend <clusterresourcename>
— Resource Groups —
Group Name Node Name State Suspended
———- ——— —– ———
Group: xxxxbdb-rg xxxxbdbsunp00 Online Yes
Group: xxxxbdb-rg xxxxbdbsunp01 Offline Yes
Group: xxxfra-rg xxxxbdbsunp00 Online Yes
Group: xxxfra-rg xxxxbdbsunp01 Offline Yes
-unmount filesystems or remove them under cluster configurations it ‘ll automatically unmount them;
# scrgadm -c -j <ClusterResourcename> -x FilesystemMountPoints=”/global/X1 /global/X2″
This command which I added above ‘ll automatically unmount X3,X4 filesystem.It ‘ll update cluster configurations
-Remove SVM device;
# metaclear -s <disksetname> d56 # metaclear -s <disksetname> d155
Step 4:Remove DID device from Disk Set;
#metaset -s <disksetname> -d /dev/did/rdsk/d79 //for SVM d56 #metaset -s <disksetname> -d /dev/did/rdsk/d67 //for SVM d155 #metaset -s <disksetname> -d /dev/did/rdsk/d65 //for SVM d155
-Check diskset If you removed them;
#metaset -s <disksetname>|egrep ‘d79|d67|d65’
-Check Metadb if you removed from metadb to;
#metadb -s <disksetname>
-If there is a configuration for the device which you removed then use;
#metadb -s <disksetname> -d /dev/did/dsk/d79
After you finished operations If your storage admin needs these luns which we removed , please check cluster quorum disk. Maybe you defined these disks for cluster quorum disk
#clq list
