Veritas Cluster Commands !!!

In this post, we are going to see veritas cluster (ha) commands.
Soon after installing VCS it is time to configure it. For configuring a cluster, we can play with GUI clicks or ha commands.

Though we can have fun playing with GUI, it is better to grab some knowledge on commands as we are admins. So let me start this.....

root@solaris:~# hares -list | grep solaris
root@solaris:~#                                               
root@solaris2:~# hagrp -list
cvm                      solaris2
cvm                      solaris
vxfen                   solaris2
vxfen                   solaris
root@solaris:~#

Till now there are no resources or service groups , so let's start creating one....
cvm and vxfen are default comes with cluster software (cvm-- cluster volume manager and vxfen -- for I/O fencing)

CMD : for a group        ---- hagrp -add <grpname>
CMD : for a resource   ---- hares -add <resource name> <resource type> <service group>

root@solaris2:~#
root@solaris2:~# hagrp -add TESTDB           ----- Adding Service Group
root@solaris2:~#
root@solaris2:~# hagrp -add TESTCI
root@solaris2:~#
root@solaris2:~# hagrp -list
TESTDB               solaris2
TESTDB               solaris
TESTCI                 solaris2
TESTCI                 solaris
cvm                      solaris2
cvm                      solaris
vxfen                   solaris2
vxfen                   solaris
root@solaris2:~#
root@solaris:~# hares -add dg_test DiskGroup TESTDB       ----- Adding resource
root@solaris:~#
root@solaris:~# hares -modify dg_test DiskGroup testdg
root@solaris:~#
root@solaris:~# hares -list | grep solaris
dg_test   solaris
root@solaris:~#

Soon after creating a resource for DiskGroup, we need a volume resource for it.

root@solaris:~#
root@solaris:~# hares -add vol_testvol  Volume TESTDB
root@solaris:~#
root@solaris:~# hares -modify vol_testvol Volume testvol
root@solaris:~#
root@solaris:~# hares -modify vol_testvol DiskGroup testdg
root@solaris:~#
root@solaris:~# hares -list | grep solaris
dg_test   solaris
vol_testvol  solaris
root@solaris:~#

Then a mount point,

root@solaris:~#
root@solaris:~# hares -add mnt_test Mount TESTDB
root@solaris:~#
root@solaris:~# hares -modify mnt_test MountPoint test
root@solaris:~#
root@solaris:~# hares -modify mnt_test BlockDevice /dev/vx/dsk/testdg/testvol
root@solaris:~#
root@solaris:~# hares -modify mnt_test FSType vxfs
root@solaris:~#
root@solaris:~# hares -modify mnt_test FsckOpt -y
root@solaris:~#
root@solaris:~# hares -list | grep solaris
dg_test   solaris
vol_testvol  solaris
mnt_test   solaris
root@solaris2:~#

We have to create resources only on one node they will be visible for both the nodes.

root@solaris:~# hares -list
dg_test   solaris
dg_test   solaris2
vol_testvol  solaris
vol_testvol  solaris2
mnt_test   solaris
mnt_test   solaris2
root@solaris:~#

Now let us go with creating a resource for ORACLE, so that we can switch oracle between the nodes and make it started just with clicks...
oracle comes with different attributes like sid(system id), owner, oracle_home, profile_file.

root@solaris:~# hares -add test_ora Oracle TESTDB
root@solaris:~#
root@solaris:~# hares -modify test_ora Sid SID
root@solaris:~#
root@solaris:~# hares -modify test_ora Owner oraSID
root@solaris:~#
root@solaris:~# hares -modify test_ora Home /oracle/SID/112_64
root@solaris:~#
root@solaris:~# hares -modify test_ora Pfile /oracle/SID/112_64/dbs/initSID.ora
root@solaris:~#

Listener is the thing which acts as a mediator for SAP and DB, hence we need a resource for this too...

root@solaris:~# hares -add  LSNR_test Netlsnr TESTDB
root@solaris:~#
root@solaris:~# hares -modify LSNR_test Owner oraSID
root@solaris:~#
root@solaris:~# hares -modify LSNR_test Home /oracle/SID/112_64
root@solaris:~#

As we use a IP for DB and SAP server, we need a IP resource for both so that we can switch DB or CI to any node. For IP, we need a NIC resource too..... Attributes like IP, Netmask and NIC.

root@solaris:~# hares -add NIC_test NIC TESTDB
root@solaris:~#
root@solaris:~# hares -modify NIC_test Device ipmp0
root@solaris:~#
root@solaris:~# hares -add VIP_test IP TESTDB
root@solaris:~#
root@solaris:~# hares -modify VIP_test Device ipmp0
root@solaris:~#
root@solaris:~# hares -modify VIP_test Address 10.20.10.21
root@solaris:~#
root@solaris:~# hares -modify VIP_test NetMask 255.255.255.0

Our ORACLE resource is ready, now let us gor SAP resource....
For SAP we need attributes like Instance type & name, sap_admin, sap_sid, profile_file .......

root@solaris:~# hares -add SID_SAP SAPNW04 TESTCI
root@solaris:~#
root@solaris:~# hares -modify SID_SAP InstType ENQUEUE 
root@solaris:~#
root@solaris:~# hares -modify SID_SAP EnvFile /home/SIDadm/.cshrc
root@solaris:~#
root@solaris:~# hares -modify SID_SAP InstName 00
root@solaris:~#
root@solaris:~# hares -modify SID_SAP ProcMon ms en pr
root@solaris:~#
root@solaris:~# hares -modify SID_SAP SAPAdmin SIDadm
root@solaris:~#
root@solaris:~# hares -modify SID_SAP SAPSID SID
root@solaris:~#
root@solaris:~# hares -modify SID_SAP StartProfile /usr/sap/SID/SYS/profile/pfl.flname

Similarly we have to create resources like NFSrestart , share and proxy if we are sharing any filesystems from a node....

root@solaris:~# hares -add  NFSrestart_test  NFSRestart TESTCI
root@solaris:~#
root@solaris:~# hares -modify NFSrestart_test LocksPathName /sapmnt/SID/nfslocks
root@solaris:~#

Proxy helps cluster to wait until the target resource is online.
Example : Will be helpful when we use nfs share. It means unless until the filesystem gets mounted on the node it cannot be shared. This scenario is handled by proxy.

root@solaris:~# hares -add  proxy_test  Proxy TESTCI
root@solaris:~#
root@solaris:~# hares -modify proxy_test TargetResName mnt_test1
root@solaris:~#
root@solaris:~# hares -add  share_test  Share TESTCI
root@solaris:~#
root@solaris:~# hares -modify share_test PathName /test1
root@solaris:~#
root@solaris:~# hares -modify share_test Options -o anon=0
root@solaris:~#

In a cluster dependency is the main thing and will be achieved by proper linking. We need to create a link between any two resources carefully and thus we define the relationship of child and parent resource respectively.

CMD : hares -link <parent res> <child res>

hares -link dg_test vol_testvol
hares -link vol_testvol mnt_test

similarly to unlink this relation :

CMD : hares -unlink <parent group> <child group>

If we want to specify dependency while any service group getting online, we can use links.
           ****    For example, we will create service group for both SAP and DB. But DB should be started before CI so in such cases, we can use link to create dependency which states unless and until DB service group is online SAP service group should not be online.

CMD :  hagrp -link <parent group> <child group>

hagrp -link TESTDB TESTCI

Similarly some more useful and regular commands are,

To bring services online and offline

hagrp -online service_group -sys system_name
hagrp -offline service_group -sys system_name

For Freezing/unfreezing service groups

hagrp -freeze group_name [-persistent]
hagrp -unfreeze group_name [-persistent]

To switch services between any nodes of cluster

hagrp -switch group_name -clus <cluster>

To bring resources online and offline

hares -online resource_name -sys system_name
hares -offline resource_name -sys system_name

Since cluster is a ocean, we have so many other commands too. But it is always easier if we go with GUI option. This post gives some basic idea regarding commands which were internally run when we click and perform some action in GUI.

################################################################################