For HACMP additional file sets are required in addition to the BOS installation file sets.
From the BOS Installation media select all “ rsct” file sets except rsct.vsd.* file sets( use smitty installp command). Then install the following file sets from HACMP Media. The file sets to be selected are
- cluster.adt.es
- cluster.es
- cluster.es.cfs
- cluster.es.cspoc
5. cluster.license
The HACMP Configuration can be done via SMIT. To start the configuration run the command # smitty hacmp and the steps are shown below in screen shots. But before starting the configuration ensure to add the boot and service IP entries in the /etc/hosts file in both the nodes that are participating in the HACMP cluster. Also configure the network in both nodes with the Boot IP and ensure the connectivity via the boot IP in both the nodes (do a ping test).
☻ First ensure that you have HACMP software/DUMP and all the required patches published on the net for the HACMP S/W
☻ First check following settings:
HIGH Watermark for pending write I/Os per File (maxpout)
# lsattr –El sys0 | grep maxpout for cluster it should be 32
LOW Watermark for pending write I/Os per File (minpout)
# lsattr –El sys0 |grep minpout for cluster it should be 24
To change the above settings:
# smitty system
Change/Show Characteristics of Operating System
and change the below parameters:
☻ Ensure that all the File Systems in the cluster are not getting mounted automatically.
# chfs –a mount=no /filesystem
For All the FS “mount=no”
☻ Also the VGs are not varied on automatically.
# chvg -a n –Q n vgname
# ls –l /dev/*vg
# chdev –l <LUN No.> –a pv=yes
§ For all the VGs autovaryon=no
§ After changing the VG characteristics in node1 unmount the file systems and varyoff the VGs from node1 and import them in the node2
# importvg –y <VG name> -V <major number> <the LUN used for the VG>
Ø MAKE SURE THAT MAJOR NO OF THE VGS IS SAME IN BOTH THE NODES.
Ø Also make sure that the file systems are not auto mounted and VGs are not auto varyon.
# chfs –a mount=no <file system name>
# chvg –a n –Q n < VG name>
☻ Try mounting of the File Systems.
☻ Then varyoff the VGs
☻ To configure Ether-channel do the following:
# ifconfig en0 down
# ifconfig en0 detach
# rmdev –dl en0
# rmdev dl et0
# rmdev –dl ent0
# cfgmgr
# smitty etherchannel
☻ Then check and remove the odm-entry as follows:
# odmget CuAt |grep <gate way>
if it shows any value like
value = “ net, -hopcount,0,,0,<gateway>”
then delete the above entry using the below command
# odmdelete –o CuAt –q value=” net, -hopcount,0,,0,<gateway>”
Again check the for the above false entry in ODM.
☻ Configure en4 with the boot IP ADDRESS and make sure that /etc/hosts is updated with the following entries in both the nodes.
o <boot ip address1> hostname1_boot
o <cluster ip Address1> hostname1_svc hostname1
o <boot ip address2> hostname2_boot
o <cluster ip Address2> hostname2_svc hostname2
☻ Add scripts to the /usr/es/sbin/cluster/
# mkdir scripts
o For HOST1(We can add scripts to below mentioned files)
# vi <hostname1>.start
route add 0.0.0.0 <gate way>
# touch <hostname1>.stop
# touch <hostname2>.start
# touch <hostname2>.stop
o For HOST2
# vi <hostname2>.start
route add 0.0.0.0 <gate way>
# touch <hostname2>.stop
# touch <hostname1>.start
# touch <hostname1>.stop
o On both the nodes
# chmod 755 *.start
# chmod 755 *.stop
☻ Configure Etherchannel. Make sure that both the ports are showing the link as UP reboot the System to make sure that the old IPs are removed completely , otherwise make the port down/detach and remove it.
# ifconfig en0 down
# ifconfig en0 detach
# rmdev –dl en0
# rmdev –dl et0
# rmdev –dl ent0
# cfgmgr
and configure etherchannel
☻ Install the HACMP software through either HACMP CD or DUMP. Also ensure that you have all the patched required for the HACMP with you. Once the HACMP S/W is installed apply all the PATCHES.
# installp –agxc <path where the dump/patches are copied> all
OR
o Use smitty installp menus
☻ Take a reboot
# shutdown –Fr
After reboot confirm all the things are done perfectly. Now the actual cluster configuration starts
☻ smitty hacmp
o Extended Configuration
o Extended Topology Configuration
o Configure an HACMP cluster
o Add/Change/Show an HACMP Cluster
Enter the cluster name
example
☻ smitty hacmp
o Extended Configuration
o Extended Topology Configuration
o Configure HACMP Nodes
o Add a Node to the HACMP Cluster
For Node 1
example
Enter the nodename1 (hostname) and the communication path to node( above mentioned name --hostname1_boot )
For Node 2
Example
Enter the nodename2 (hostname) and the communication path to node( above mentioned name –hostname2_boot )
☻ smitty hacmp
o Extended Configuration
o Discover HACMP-related Information from configured Nodes
Configuring Heart Beat on Non IBM Storage..
☻ smitty hacmp
o Extended Configuration
o Extended Topology Configuration
o Configure HACMP Communication Interface devices
o Add Communication Interface/Devices
o Add Discovered Communication interface and Devices
o Communication Interfaces
Example
Select the two discovered communication path to node
☻ smitty hacmp
o Extended Configuration
o Extended Resource Configuration
o HACMP Extended Resources Configuration
o Configure HACMP Service IP Labels/Addresses
o Add a Service IP Label/Address
o Configurable on Multiple Nodes.
Example
☻ smitty hacmp
o Extended Configuration
o Extended Resource Configuration
o HACMP Extended Resource Configuration
o Configure HACMP Application
o Configure HACMP Application Servers
o Add An Application Server
Node1
Add Server name and path of the scripts
(You have to mention the path for the scripts for the hostname1.start/stop for both nodes)
Node2
Server name and path of the scripts
(You have to mention the path for the scripts for the hostname1.start/stop for both nodes)
☻ smitty hacmp
o Extended configuration
o Extended Resource Configuration
o HACMP Extended Resource Group Configuration
o Add a Resource Group
§ (Here for the host2 to host1 failover type the host2 host1 sequence manually in the Participating Nodes(Default Node Priority)
§ option )
For node2(Enter the group name2 ,the priority and the failover policies)
example
For node2(Enter the group name2 ,the priority and the failover policies)
For concurrent volume groups
☻ smitty hacmp
o Extended Configuration
o Extended Resource Configuration
o HACMP Extended Resource Group Configuration
o Change /Show resources and Attributes for a Recovery Group
Here three options have to be filled :
a. Service IP labels/Addresses
b. Application Server
c. Volume Groups
For node1
For Node2
For concurrent Volume group
☻ smitty hacmp
o Extended Configuration
o Extended Verification and Synchronization
For disk heart beats
☻ smitty hacmp
o Extended Configuration
o Extended Topology Configuration
o Configure HACMP Communication Interface devices
o Add Communication Interface/Devices
o Add Discovered Communication interface and Devices
Select Communication Devices
☻ smitty hacmp
o Extended Configuration
o Extended Verification and Synchronization
If the cluster sync fails then do as follows:
Go to /usr/es/sbin/cluster/diag
Cp ver_get_diskhb_info ver_get_diskhb_info_org
Vi ver_get_diskhb_info
search for mode and change the below entry
Original entry : mode=`/usr/sbin/lqueryvg –p $line –X >/dev/null 2>&1`
Modified entry : mode=`/usr/sbin/lqueryvg –p $line –X 2>/dev/null`
Then sync the cluster.
Start the cluster.
If hacmpvg is not showing concurrent in lspv then create new resource group and add resource vg in resource group and do the necessary changes shown below…..
No comments:
Post a Comment