To migrate from a 1.2 system to a 2.1.x system, you must examine your ha.conf file in order to determine how to define the equivalent parameters in the 2.1.x cluster database.
The following sections show upgrade examples for the following tasks:
 | Note: The examples in this paper use the cmgr command in prompting mode (-p) and reflect the FailSafe 2.1.5 release. For details about cmgr, see the IRIS FailSafe Version 2 Administrator's Guide. |
The following example shows node definition in the 1.2 ha.conf file. Parameters that you must use when configuring a 2.1. x system are indicated in bold:
Node node1 { interface node1-fxd { name = rns0 ip-address = 54.3.252.6 netmask = 255.255.255.0 broadcast-addr = 54.3.252.6 } heartbeat { hb-private-ipname = 192.0.2.3 hb-public-ipname = 54.3.252.6 hb-probe-time = 6 hb-timeout = 6 hb-lost-count = 4 } reset-tty = /dev/ttyd2 sys-ctlr-type = MSC } |
In this configuration example, you will use the following values when you define the same node in 2.1.x:
Node name: node1
Primary network interface: node1
Type of system controller: msc
System control device name: /dev/ttyd2
Control networks: 192.0.2.3, 54.3.252.6
Use the following cmgr command to use these values to define a node in 2.1.x.
| Note: There are additional parameters you must specify when you define this node. |
cmgr> define node node1 Enter commands, you may enter "done" or "cancel" at any time to exit Hostname[optional]? node1 Is this a FailSafe node <true|false> ? true Is this a CXFS node <true|false> ? false Operating System <IRIX|Linux32|Linux64|AIX|HPUX|Solaris|Windows> ? Irix Node Function <server_admin|client_admin> ? server_admin Node ID[optional] ? Partition ID[optional] ? (0) Reset type <powerCycle> ? (powerCycle) Do you wish to define system controller info[y/n]:y Sysctrl Type <msc|mmsc|l2|l1>? (msc) Sysctrl Password [optional]? ( ) Sysctrl Status <enabled|disabled>? enabled Sysctrl Owner? node2 Sysctrl Device? /dev/ttyd2 Sysctrl Owner Type <tty> [tty]? Number of Network interfaces [1]? 2 NIC 1 - IP Address? 192.0.2.3 NIC 1 - Heartbeat HB (use network for heartbeats) <true|false>? true NIC 1 - (use network for control messages) <true|false>? true NIC 1 - Priority <1,2,...>? 1 NIC 2 - IP Address? 192.0.2.4 NIC 2 - Heartbeat HB (use network for heartbeats) <true|false>? true NIC 2 - (use network for control messages) <true|false>? true NIC 2 - Priority <1,2,...>? 2 |
For information on setting monitoring values in 2.1. x, see “Setting HA Parameters”.
Although 1.2 does not require the definition of clusters, you specify a parameter in the ha.conf file that 2.1.x uses in its cluster definition: the e-mail address to use to notify the system administrator when problems occur in the cluster.
The ha.conf file includes the following:
system configuration
{
mail-dest-addr = root@localhost
...
} |
When you define a cluster in 2.1.x, you can use this as the e-mail address for problem notification.
There are other things you must provide in addition to this parameter when you define a 2.1.x cluster, such as the e-mail program to use for this notification and, of course, the nodes to include in the cluster.
Use the following cmgr command to define a cluster:
cmgr> define cluster apache-cluster Enter commands, you may enter "done" or "cancel" at any time to exit Is this a FailSafe cluster <true|false> ? true Is this a CXFS cluster <true|false> ? false Cluster Notify Cmd [optional] ? Cluster Notify Address [optional] ? Cluster HA mode <normal|experimental> [optional] ? No nodes in cluster apache-cluster Add nodes to or remove nodes from cluster clustername Enter "done" when completed or "cancel" to abort clustername ? add node node1 clustername ? add node node2 ... clustername ? done Successfully defined cluster apache-cluster |
The following example shows the sections of a 1.2 ha.conf file that are used to set monitoring and timeout values. Parameters that you must use when configuring a 2.1.x system are indicated in bold.
system-configuration
{
pwrfail = true
...
}
Node node1
{
...
heartbeat
{
hb-private-ipname = 192.0.2.3
hb-public-ipname = 54.3.252.6
hb-probe-time = 6
hb-timeout = 6
hb-lost-count = 4
}
...
} |
As this ha.conf node-definition shows, in 1.2 you defined hb-probe-time, hb-timeout, and hb-lost-count parameters to set the values that determined how often to send monitoring messages and the length of time without a response that would indicate a failure. The 2.1. x release uses a different method for monitoring the nodes in a cluster than 1.2 uses, sending out continuous messages to the other nodes in a cluster and, in turn, maintaining continuous monitoring of the messages the other nodes are sending.
Because of the different monitoring methods between the two systems, there is no one-to-one correspondence between the values you set in the 1.2 ha.conf file and the timeout and heartbeat intervals you set in 2.1.x when you set FailSafe HA parameters. However, if you wish to maintain approximately the same time interval before which your system determines that failure has occurred, you can use the following formula to determine the value to which you should set your node timeout interval:
node_timeout = (probetime + timeout) * lostcount |
This formula should account for the same total node-to-node communication time.
All 2.1.x timeouts are in milliseconds, and can be changed when 2.1.x is running. Timeouts can be specified for the cluster for a specific node in the cluster.
There is no long-timeout value in 2.1.x. The long-timeout value equivalent is set with the resource type start and stop timeouts. The resource type start, monitor, and stop timeouts can be changed using the GUI or cmgr.
Use the following cmgr command to modify the HA parameters for node1 in 2.1.x:
| Note: Node-specific cluster database information is not replicated to other nodes in the cluster. Therefore, you must change node-specific information when running cmgr on that node. You must execute cmgr -p on node1 in the following example. |
cmgr> modify ha_parameters on node node1 in cluster apache-cluster Enter commands, when finished enter either "done" or "cancel" Node Timeout (60000) ? 24000 Heartbeat Period (1000)? 6000 Node wait time (12000)? Tie Breaker Node? Run Powerfail (true)? Successfully modified ha_parameters |
The following example shows a volume definition in the 1.2 ha.conf file. Parameters that you must use when configuring the same volume as a volume resource in a 2.1.x system are indicated in bold:
volume apache-vol
{
server-node = node1
backup-node = node2
devname = apache-vol
devname-owner = root
devname-group = sys
devname-mode = 600
} |
In this configuration example, you will use the following values when you define the same volume in 2.1.x:
Volume name: apache-vol
User name of device file owner: root
Group name of device file: sys
Device file permissions: 600
To create an XLV volume resource, use the following cmgr command and the default values for each argument:
cmgr> define resource apache-vol of resource_type volume in cluster apache-cluster Enter commands, when finished enter either "done" or "cancel" devname-group [optional] ? (sys) devname-owner [optional] ? (root) devname-mode [optional] ? (600) No resource type dependencies Add dependencies to or remove from resource apache-vol: Enter "done" when completed or "cancel" to abort apache-vol ? done |
The following example shows an XFS filesystem definition in the 1.2 ha.conf file. Parameters that you must use when configuring the same filesystem as a filesystem resource in a 2.1.x system are indicated in bold:
filesystem apache-fs
{
mount-point = /apache-fs
mount-info
{
fs-type = xfs
volume-name = apache-vol
mode = rw, noauto
}
} |
In this configuration example, you will use the following values when you define the same filesystem in 2.1.x:
Resource name (mount point): /apache-vol
XLV volume: apache-vol
Mount options: rw, noauto
To create a filesystem resource, use the following cmgr command:
cmgr> define resource /apache-fs of resource_type filesystem in cluster apache-cluster Enter commands, you may enter "done" or "cancel" at any time to exit volume-name [optional] ? apache-vol mount-options [optional] ? (rw) rw,noauto kill-nfsds-before-umount [optional] ? (true) monitoring-level [optional] ? (2) Resource type dependencies to add or remove: Resource dependency type - 1: volume Add dependencies to or remove from resource /lori-fs: Enter "done" when completed or "cancel" to abort /apache-fs ? done Successfully defined resource /apache-fs |
The following example shows an IP address definition in the 1.2 ha.conf file. Parameters that you must use when configuring the same IP address as a highly available resource in a 2.1. x system are indicated in bold:
interface-pair FDDI_1
{
primary-interface = node-fxd
secondary-interface = node2-fxd
re-mac = false
netmask = 0xffffff00
broadcast-addr = 54.3.252.255
ip-aliases = ( 54.3.252.7 )
} |
In this configuration example, you will use the following values when you define the same IP Address in 2.1. x:
Resource name: 54.3.252.7
Broadcast address: 54.3.252.255
Network mask: 0xffffff00
To create an IP address resource, use the following cmgr commands:
cmgr> define resource 54.3.252.7 of resource_type IP_address in cluster apache-cluster Enter commands, you may enter "done" or "cancel" at any time to exit NetworkMask [optional] ? 0xffffff00 interfaces [optional] ? rns0 BroadcastAddress [optional] ? 54.3.252.255 No resource type dependencies Add dependencies to or remove from resource 54.3.252.7: Enter "done" when completed or "cancel" to abort 54.3.252.7 ? done Successfully defined resource 54.3.252.7 |