Академический Документы
Профессиональный Документы
Культура Документы
html
Contents
AIX
1. Introduction
BSD 2. STEP 1: Additional Boot Addresses
3. STEP 2: Import LUNs of shared volume groups into new node
HP-UX 4. STEP 3: Add the new node to the cluster
5. STEP 4: Restart cluster services on new node
Linux 6. STEP 5: Add new IP addresses to the cluster configuration
7. STEP 6: Start HACMP on new node
Solaris
8. STEP 7: Remove Dependencies between Resource Groups
9. STEP 8: Add new node to all resource groups
Others
10. STEP 9: Restore dependencies between Resource Groups
Un*x 11. STEP 10: Final synchronization of the cluster
A. Related Information
Topics
Introduction
Guestbook This article describes how to add a third node hoppy to an HACMP cluster. We find the following situation:
Conclusion: barney and betty are members of two resource groups, but all resources are currently active on
barney.
Note: Usually the boot addresses are not equal to the host specific ip addresses of the interfaces (such as
the hostname's ip address). They become persistent addresses instead. However, in a virtualized
environment we could use the host specific IP addresses as boot addresses and could avoid the need for
We add the new node's boot and persistent addresses to /etc/hosts on all nodes. The old node's IP
addresses as well as all cluster service addresses have to be added to the new node's /etc/hosts.
All storage assigned to the resource groups RES_GRP_01 and RES_GRP_02 has to be visible to the new
node. Be sure to set the reserve_policy to no_reserve for all shared storage on the new node:
hoppy# cfgmgr
hoppy# chdev -a reserve_policy=no_reserve hdiskN1
hoppy# chdev -a reserve_policy=no_reserve hdiskN2
hoppy# chdev -a reserve_policy=no_reserve hdiskN3
.
.
.
Now we are ready to import all shared Volume Groups into the new node.
Get information about major number, Volume Group, and PVID from existing nodes.
Now we are ready to add hoppy as new node to the cluster. We do this on one of the existing nodes.
[Entry Fields]
* Node Name [hoppy]
Communication Path to Node [<boot address of new node
To be sure we check /usr/es/sbin/cluster/etc/rhosts on all nodes -> all cluster addresses should be in it!
At this point the cluster services on the new node can be restarted:
[Entry Fields]
* IP Label/Address [<new boot address
* Network Type ether
* Network Name (selected network)
* Node Name []
Network Interface []
When we are done with the boot addresses we can put the new node's persistent addresses to the
interfaces:
[Entry Fields]
* Node Name hoppy
* Network Name [(selected network)]
* Node IP Label/Address [<persistent IP address
Netmask(IPv4)/Prefix Length(IPv6) []
If our persistent address is in a different network than the boot address and there are routes through the
interface (what is the classical setup) they have to be defined in the ODM on hoppy. For a default route to a
gateway 111.111.111.1 the command would be
Any online on same node dependencies have to be removed before we are able to add the new node to
resource groups:
Select resource groups one by one and add the new node to the end of the list of Participating Nodes:
[Entry Fields]
Resource Group Name (Resource Group)
New Resource Group Name []
Participating Nodes (Default Node Priority) [barney betty hoppy
Note: Really add the new node to the end of the list. Don't use the F4 selection, because the order might be
wrong! To be sure we verify the node order in the resource group with
barney# clRGinfo
-----------------------------------------------------------------------------
Group Name Group State Node
-----------------------------------------------------------------------------
RES_GRP_01 ONLINE barney
OFFLINE betty
OFFLINE hoppy
A. Related Information
AIX 7.1 Information Center > PowerHA SystemMirror > HACMP Version 6.1 > Adding a cluster node
to the HACMP configuration
Also on unixwerk
HACMP: Cluster Commandline
How to Add a New VG to an Active HACMP Resource Group
How to Add a New Resource Group to an Active Cluster
How to Remove a Node from an HACMP Cluster
Setup a Two-Node Cluster with HACMP