Thursday, 14 July 2011

How to configure HACMP V5.x with IPAT(like HACMP V4)

How to configure HACMP V5.x with IPAT(like HACMP V4)
Cluster Topology Configuration.

NOTE: Underscore ( _ ) is the only valid special character for HACMP objects names
      First character must not be numeric.

From one node, you will execute the following steps
    # smit hacmp
             Extended Configuration
               Extended Topology Configuration                
         Configure Cluster
         Configure Nodes
        Configure IP Networks
                               Configure Communication Interfaces on these IP networks
                        Configure non-IP Networks (serial rs232 or disk heartbeat)
                        Define Communication Devices on these non-IP networks
                       Configure persistent IP addresses for network(s) with boot or service
          Synchronize configuration.    It will also setup persistent IP aliases

 1) Nodes prerequisites

- Network configuration on EACH node:
. interfaces on their “boot”  and  “standby” address, each one on a different subnet    .   . hostname corresponding to future “persistent” address: smit mkhostname
 /etc/hosts must contain all nodes addresses, persistent and application IP aliases
.  /.rhosts should be empty, except for Bull ClusterWatch (corrected in future PTF).
- on each node file  /usr/es/sbin/cluster/etc/rhosts must contain all the “boot” addresses of all nodes and all HACMP networks, for example:
node1_boot1
node2_boot1

- Check also that the following daemon is running :
# lssrc –g clcomdES
     clcomdES         clcomdES         16310   active
  otherwise start it with     startsrsc –s clcomdES

- godm TCP/IP service must be active ( refresh –s inetd  on all nodes)

- Volume Groups imported on all nodes, not automatically activable at boot time,
   If AIX 5.2, they must be concurrent capable (bos.clvm.enh fileset installed)

1) Configure Cluster
    # smit hacmp
      Extended Configuration
          Extended Topology Configuration
         Configure an HACMP Cluster
          Add/Change/Show an HACMP Cluster
        Assign a unique Cluster Name (< 32 characters)


2) Configure Nodes
      # smit hacmp
      Extended Configuration
          Extended Topology Configuration  
           Configure HACMP Nodes
         Add a Node to the HACMP Cluster
     Give Node Name, each node has a unique name (<32 characters)
        Communication Path to Node: Press F4 and select an IP label (e.g.: boot address)
        Repeat this step for all nodes.

3) Configure IP Networks  
    # smit hacmp
      Extended Configuration
          Extended Topology Configuration
                   Configure HACMP Networks
                      Add a Network to the HACMP Cluster
Go to    # Pre-defined IP-based Network Types
select network type (e.g.: ether)
check  that Netmask is correct
It is mandatory to modify the following field:
* Enable IP Address Takeover via IP Aliases          [No]          
Repeat this step for all IP networks (interconnect, administration).      

then Configure Communication Interfaces on these IP networks
    # smit hacmp
      Extended Configuration
          Extended Topology Configuration
  Configure HACMP Communication Interfaces/Devices
             Add Communication Interfaces/Devices  
                               Add Pre-defined Communication Interfaces and Devices
                                   Communication Interfaces
Select a previously defined  network
give IP Label and Node Name
Repeat this step for all  “boot and standby” addresses of all IP networks and of all nodes. E.g.: node1_boot, node1_stby, node2_boot, node2_stby
                     

4) Configure non-IP network (serial rs232 or disk heartbeat)
# smit hacmp
Extended Configuration
           Extended Topology Configuration
    Configure HACMP Networks
               Add a Network to the HACMP Cluster
     Go to   # Pre-defined Serial Device Types
     Select  rs232 or diskhb
This will create a network with a default name like net_rs232_0<x>

then Define Communication Devices on these non-IP networks
     # smit hacmp
      Extended Configuration
          Extended Topology Configuration
              Configure HACMP Communication Interfaces/Devices
                   Add Communication Interfaces/Devices
                    Add Pre-defined Communication Interfaces and Devices
                Communication Devices
          Select a Device Network then give:
                    Device Name:    symbolic name eg.  node1_tty
        Device Path  :    /dev/tty<x>  or    /dev/hdiskpower<x>   (or  /dev/hdisk<x>)
                     disk must belong to an  AIX 5.2 enhanced concurrent VG
         Node Name   :     name of one node
       
Repeat this step to define the Communication Device of the other node(s) …


5) Configure persistent IP addresses for network(s) where an initial  “boot” address which will be replaced by a “service” address
    # smit hacmp
       Extended Configuration
              Extended Topology Configuration
                 Configure HACMP Persistent Node IP Label / Addresses
               Add a Persistent Node IP Label / Address
       Select a Node and give:
      * Network Name            [ ]          Press  F4
    * Node IP Label / Address    [ ]          Should usually correspond to node hostname
                             or Press  F4 to select from /etc/hosts

Repeat this step for the other interface(s) on other node(s)  …

6) Synchronize configuration. It will also setup persistent addresses
# smit hacmp
     Extended Configuration
         Extended Verification and Synchronization


 # netstat –i   shows persistent IP addresses on the same interfaces as boot  addresses.

Cluster Resources Configuration
From one node, you will execute the following steps:
    # smit hacmp
Extended Configuration
    Extended Resource Configuration
                     Configure HACMP Service IP Labels/Addresses (srv address)
        Define Application Servers
        Define Resource Groups
        Configure Resources for each Resource Group  
    Customization
        Synchronize Cluster Resources
    Start HACMP on nodes

1) Configure HACMP Service IP Labels/Addresses
# smit hacmp
Extended Configuration
             Extended Resource Configuration
                 HACMP Extended Resources Configuration
                       Configure HACMP Service IP Labels/Addresses
                          Add a Service IP Label/Address
Select:                           Configurable on Multiple Nodes
Select a network name
Give interface service IP label or address (on the same subnet as “boot” address)
You shouldn’t  specify an Alternate HW Address to accompany IP Label/Address
because AIX 5L does “gratuitous ARP”  update.
Don’t  specify an Alternate HW Address for Ethernet Gigabit adapters

Repeat this step for ALL service IP labels.

2) Define Application Servers
    # smit hacmp
Extended Configuration
               Extended Resource Configuration
                   HACMP Extended Resources Configuration
                        Configure HACMP Application Servers
                           Add an Application Server
Give:
    - Server name (symbolic name for the resource)
    - Start and Stop scripts full pathnames  (must exist on ALL NODES, in non-shared
       filesystems).

Repeat this step for ALL Application Servers.

3) Define Resource Groups
    # smit hacmp
Extended Configuration
               Extended Resource Configuration
                  HACMP Extended Resource Group Configuration
                 Add a Resource Group
Select the Resource Group Management Policy: cascading, rotating, concurrent or custom
Give Resource Group Name,
Inter-Site Management Policy (leave default  ignore)
Give the list of Participating Node Names: for cascading the order defines the priority.

Repeat this step for all Resource Groups

4) Configure Resources for each Resource Group
           # smit hacmp
Extended Configuration
             Extended Resource Configuration
                 HACMP Extended Resource Group Configuration
               Change/Show Resources and Attributes for a Resource Group  
Select a Resource Group name
Define the resources belonging to the resource group (separated with space):
Inactive Takeover Applied   (true if you want first starting node to take all resources)
Cascading Without Fallback Enabled (true to decide when fallback will occur
                                 recommended if HACMP  Cluster Services started in /etc/inittab)
Application Servers
Service IP Labels / Addresses  (Give ALL Service IP labels separated by space,
                     in case of several networks)
    Volume Groups   give the name(s) separated by space
Use forced varyon of volume groups, if necessary (true for AIX mirrored VGs)
    Filesystems (empty is All)   leave empty
Filesystems Recovery Method    leave sequential if nested recursive mount points
otherwise, parallel when many
Filesystems mounted before IP configured  (TRUE for NFS (cross)-mount)
    Filesystems/Directories to Export     (must belong to Volume Groups listed before)
    Filesystems/Directories to NFS mount (must belong to previous Export list)
Network for NFS mount
 
 Repeat this step for all Resource Groups


Synchronize Cluster Resources (everytime you change the configuration)

# smit hacmp
      Extended Configuration
Extended Verification and Synchronization
which allows to:
     * Verify, Synchronize or Both                         [Both]      
      Force synchronization if verification fails?        [No]        
* Verify changes only?     [No]        
* Logging               [Standard / Verbose]
In case of problem, select Verbose Logging and look to log files:
     /var/hacmp/clverify/clverify.log  or /var/hacmp/clverify/…

Now you can start HACMP on all nodes (several nodes at the same time)

# smit clstart
* Start now, on system restart or both          now                  
  Start Cluster Services on these nodes    [node1]    you can specify several nodes    

No comments:

Post a Comment

Twitter Bird Gadget