What is a communication path?

LifeKeeper configures clusters by connecting multiple nodes to each other. Availability can be maintained by switching to resources on another node in the cluster in the event that a node fails. To achieve this, it is necessary to configure paths for communication between nodes in advance. These paths are called “communication paths” in LifeKeeper.

When communication paths are configured, each node sends a signal called a “heartbeat”. The heartbeat indicates that LifeKeeper is working properly. If the heartbeat is not received, the node is considered to have failed. SIOS recommends connecting cluster nodes with two or more communication paths so that in the case of a network failure, the nodes will still be able to receive heartbeats over an alternate path.

Communication paths are used for LifeKeeper internal communication, in addition to heartbeats. A remote LifeKeeper node can be managed over the network as long as communication paths are connected between the nodes.

This guide describes the steps to create and delete a single communication path between two nodes.

Configuration

The commands and other information in this guide are based on the following diagram.

Before you begin, please check that the following conditions are met in your environment:

  • TCP Port 7365 is available for the communication path
    • The firewall allows communication or the firewall is disabled
  • There are two machines running LifeKeeper

Next, record information about your systems that will be required in order to execute the CLI commands.

  1. Available IP address on each system
# ip address | grep 'inet'
    inet 127.0.0.1/8 scope host lo
    inet6 ::1/128 scope host
    inet 10.1.6.88/16 brd 10.1.255.255 scope global noprefixroute ens192
    inet6 fe80::a633:2758:4976:b42/64 scope link noprefixroute
    inet 192.168.129.2/24 brd 192.168.129.255 scope global noprefixroute ens224
    inet6 fe80::34a4:417f:b58b:dc15/64 scope link noprefixroute
    inet 192.168.122.1/24 brd 192.168.122.255 scope global virbr0
  1. Host name of each system
# hostname
target1

Creating a LifeKeeper Communication Path

Note: The command for creating a communication path creates the path in only one direction. Therefore, in order for the communication path to be able to communicate in both directions, you need to run the command on both systems.

Perform the following steps on target1

  1. Make sure that a communication path is not connected.
[target1]# lkcli status -q
LOCAL    TAG   ID    STATE     PRIO  PRIMARY
  1. Create a communication path from target1 to target2.
[target1]# lkcli commpath create --laddr 10.1.6.88 --raddr 10.1.6.89 --dest target2
Performing commpath 'target2:10.1.6.88/10.1.6.89' create...
Commpath 'target2:10.1.6.88/10.1.6.89' created successful.

Command Argument

Item Input Value
--laddr IP address on the local node to connect from
--raddr IP address on the remote node to connect to
--dest Host name of the remote node to connect to
  1. Verify that a communication path to target2 has been created. (Note: At this point, the STATE is DEAD because only a one-way connection is established.)
[target1]# lkcli status -q
LOCAL    TAG   ID    STATE     PRIO  PRIMARY
 
MACHINE  NETWORK ADDRESSES/DEVICE     STATE     PRIO
target2  TCP     10.1.6.88/10.1.6.89  DEAD         1

Perform the following steps on target2

  1. Make sure a communication path is not connected.
[target2]# lkcli status -q
LOCAL    TAG   ID    STATE     PRIO  PRIMARY
  1. Create a communication path from target2 to target1.
[target2]# lkcli commpath create --laddr 10.1.6.89 --raddr 10.1.6.88 --dest target1
Performing commpath 'target1:10.1.6.89/10.1.6.88' create...
Commpath 'target1:10.1.6.89/10.1.6.88' created successful.

Command Arguments

Item Input Value
--laddr IP address on the local node to connect from
--raddr IP address on the remote node to connect to
--dest Host name of the remote node to connect to
  1. Verify that the communication path is established. Confirm that the STATE is ALIVE.

*It may take a few seconds for the communication path to become ALIVE.

[target2]# lkcli status -q
LOCAL    TAG   ID    STATE     PRIO  PRIMARY
 
MACHINE  NETWORK ADDRESSES/DEVICE     STATE     PRIO
target1  TCP     10.1.6.89/10.1.6.88  ALIVE        1

Deleting a LifeKeeper Communication Path

Note: The command for deleting a communication path deletes the path in only one direction. Therefore, in order for the communication path to be deleted in both directions, you need to run the command on both systems.

Perform the following steps on target2

  1. Make sure that there is a connected communication path.
[target2]# lkcli status -q
LOCAL    TAG   ID    STATE     PRIO  PRIMARY
 
MACHINE  NETWORK ADDRESSES/DEVICE     STATE     PRIO
target1  TCP     10.1.6.89/10.1.6.88  ALIVE        1
  1. Delete the communication path from target2 to target1.
[target2]# lkcli commpath delete --laddr 10.1.6.89 --raddr 10.1.6.88 --dest target1

Command Arguments

Item Input Value
--laddr IP address on the local node to connect from
--raddr IP address on the remote node to connect to
--dest Host name of the remote node to connect to
  1. Confirm that the communication path to target1 has been deleted.
[target2]# lkcli status -q
LOCAL    TAG   ID    STATE     PRIO  PRIMARY

Perform the following steps on target1

  1. Make sure that a communication path to target2 still exists.
[target1]# lkcli status -q
LOCAL    TAG   ID    STATE     PRIO  PRIMARY
 
MACHINE  NETWORK ADDRESSES/DEVICE     STATE     PRIO
target2  TCP     10.1.6.88/10.1.6.89  DEAD         1
  1. Delete the communication path to target2.
[target1]# lkcli commpath delete --laddr 10.1.6.88 --raddr 10.1.6.89 --dest target2
  1. Confirm that the communication path to target2 has been deleted.
[target1]# lkcli status -q
LOCAL    TAG   ID    STATE     PRIO  PRIMARY

Feedback

Thanks for your feedback.

Post your comment on this topic.

Post Comment