Chapter 1. Introduction

IRIS FailSafe 2.1.x is not a new release of the IRIS FailSafe 1.2 product. Instead, it is a new set of files and scripts that provides many additional possibilities for the size and complexity of a highly available system.

If you wish to migrate a 1.2 system to a 2.1.x system to take advantage of these features, you must upgrade your system configuration. There is no upgrade installation option to automatically upgrade 1.2 to 2.1.x.

This chapter discusses the following:

See also Chapter 2, “Configuration Examples” and Chapter 3, “Script Changes for Programmers”.

Summary of Changes

In 1.2, the unit of failover is the node. In 2.1.x, the unit of failover is the resource group. Because of this, the concepts of node failover, node failback, and node state do not apply to 2.1. x. In addition, all FailSafe scripts differ between the two releases.

Table 1-1 summarizes the differences between the 1.2 and 2.1.x releases.

Table 1-1. Differences between 1.2 and 2.1. x

FailSafe 1.2

FailSafe 2.1. x

ha.conf configuration file.

Cluster database at /var/cluster/cdb/cdb/db. The database is automatically copied to all nodes in the pool.

Much of the data contained in the 1.2 ha.conf file will be used in the 2.1. x database, but the format is completely different. You will configure the database using the FailSafe Manager graphical user interface (GUI) or the cmgr command.

Node states (standby, normal, degraded, booting or up).

Resource group states (online, offline, pending, maintenance, error).

Scripts:

giveaway, giveback

takeover, takeback

check

(no equivalent)

Scripts:

stop

start

monitor

exclusive, probe, restart

Failover script

Failover attributes

All common functions and variables are kept in the /var/ha/actions/common.vars file.

All common functions and variables are kept in the /var/cluster/ha/common_scripts/scriptlib file.

Configuration information is read using the ha_cfginfo command.

Configuration information is read using the ha_get_info() and ha_get_field() shell functions.

Software links specify application ordering.

Software links are not used for ordering.

Scripts use /sbin/sh.

Scripts use /sbin/ksh .

Scripts require configuration checksum verification.

There is no configuration checksum verification in the scripts.

Scripts require resource ownership.

Action scripts have no notion of resource ownership.

Scripts do not run in parallel.

Multiple instances of action scripts can be run at the same time.

Each service has its own log in /var/ha/logs.

Action scripts use cluster logging and all scripts log to the same file using the ha_cilog command.

There are two units of failover, one for each node in the cluster.

There is a unit of failover (a resource group) for each highly available service.


Hardware Changes

There are no hardware changes that are required when you upgrade a system to 2.1.x. A 1.2 system will be a dual-hosted storage with reset ring two-node configuration in 2.1.x.

With 2.1.x, you can test the hardware configuration with FailSafe diagnostic commands. (See IRIS FailSafe Version 2 Administrator's Guide, for instructions on using FailSafe to test the connections.) These diagnostics are not run automatically when you start FailSafe 2.1.x; you must run them manually.

You can also use the admin ping command to test the serial reset line in 2.1.x. This command replaces the ha_spng command you used with FailSafe 1.2.

The following 1.2 command tests serial reset lines:

# /usr/etc/ha_spng -i 1 -d msc -f /dev/ttyd2
# echo $status

The following 2.1.x cmgr command tests serial reset lines:

cmgr> admin ping dev_name /dev/ttyd2 of dev_type tty with sysctrl_type msc

If the crsd daemon is running, this command will not run.

See the IRIS FailSafe Version 2 Administrator's Guide for information on using cmgr commands.

Software Changes

FailSafe 2.1.x consists of a different set of files than 1.2. The 1.2 and 2.1.x software can exist on the same node, but you cannot run both versions of FailSafe at the same time.

FailSafe 1.2 contains a configuration file, ha.conf. In 2.1.x, configuration information is contained in a cluster database at /var/cluster/cdb/cdb.db that is kept in all FailSafe nodes in the pool. You configure the cluster database using the cmgr command or the GUI.


Note: If you are running 2.1.x in coexecution with CXFS, there may be some CXFS client-only nodes in the cluster that do not contain the cluster database.

The 2.1.x cluster database is automatically copied to all administration nodes in the pool. The 2.1. x configuration is kept in all administration nodes in the pool.

Configuration Changes

You must reconfigure your 1.2 system by using the 2.1.x FailSafe Manager GUI or the 2.1.x cmgr command. For information on using these administration tools, see the IRIS FailSafe Version 2 Administrator's Guide.

To update a 1.2 configuration, consider how the 1.2 configuration maps onto the concept of resource groups:

  • A dual-active 1.2 configuration contains two resource groups, one for each node.

  • An active/standby 1.2 configuration contains one resource group, consisting of an entire node (the active node).

Each resource group contains all the applications that were primary on each node and backed up by the other node.

When you configure a 2.1.x system, you perform the following steps:

  1. Add nodes to the pool.

  2. Define the cluster.

  3. Add nodes to the cluster.

  4. Set HA parameters (FailSafe 2.1. x can be started at this point, if desired).

  5. Define resources.

  6. Define failover policies.

  7. Define resource groups.

  8. Add resources to resource groups.

  9. Put resource groups online.

The guided configuration tasks in the GUI lead you through these steps.

Script Changes

All customized 1.2 scripts must be rewritten for 2.1.x. For more information, see Chapter 3, “Script Changes for Programmers”.

System Status Changes

In 1.2, you produced a display of the system status with the ha_admin -a command. In 2.1.x, you can display the system status in the following ways:

  • You can keep continuous watch on the state of a cluster using the GUI.

  • You can query the status of an individual resource group, node, or cluster using either the GUI or cmgr.

  • You can use the /var/cluster/cmgr-scripts/haStatus script provided with the cmgr command to see the status of the cluster, nodes, resources, and resource groups.

For information on performing these tasks, see the IRIS FailSafe Version 2 Administrator's Guide.

Additional 2.1.x Tasks

After you have defined your nodes, cluster, and resources, you define your resource groups, a task which has no equivalent in FailSafe 1.2. When you define a resource group, you specify the resources that will be included in the resource group and the failover policy that determines which node will take over the services of the resource group on failure.

After you have configured your system, you can start FailSafe services, as described in the IRIS FailSafe Version 2 Administrator's Guide.