Top
PRIMECLUSTER Global Link Services Configuration and AdministrationGuide 4.3Redundant Line Control Function

3.4.2 NIC switching mode

This section describes how to change the settings for NIC switching mode. After changing the settings, you need to reflect the changes in the operations following some procedures. Note that the change distribution procedures vary depending on the command used for changing the settings, and whether the settings were changed in a single system configuration (no cluster is used), or in a cluster configuration.

Distribution Procedure

hanetconfig command

Single

Cluster

IP address to be assigned for the virtual or physical interface (-i, -e)

1

2

Virtual interface (-n)

1

2

Physical interface (-t)

1

2

Name of the virtual interface monitored by the standby patrol (-t)

1

2

hanetmask command

Single

Cluster

Subnet mask (-m)

5

5

hanetparam command

Single

Cluster

Standby patrol monitoring period (-m)

4

4

The number of constant standby monitoring (-p)

4

4

Hostname resolution (-h)

4

4

hanetpoll command

Single

Cluster

a monitor-to ip address (-p)

1

2

Setting of HUB-to-HUB monitoring function (-b)

1

2

Monitoring period (-s)

3

3

The number of monitoring (-c)

3

3

Cluster switching (-f)

-

3

Link up waiting period (-p)

3

3

Link status monitoring (-l)

3

3

Network configuration of OS

Single

Cluster

Network configuration file (/etc/sysconfig/network-scripts/ifcfg-ethX, /etc/sysconfig/network), hosts file(/etc/hosts) etc.

6

6


Procedure 1

Perform one of the following procedures after changing settings.

Procedure 2

Perform one of the following procedures after changing settings.

Procedure 3

Perform one of the following procedures after changing settings.

Procedure 4

Changed settings are immediately reflected in the operations after executing the command to change settings. No distribution procedure is required.


Procedure 5

Perform one of the following procedures after changing settings.

Procedure 6

If you modified the network configuration file for the operating system, you must reboot the system instead of manually restarting the network service (/etc/init.d/network restart, service network restart).

Changing Procedure

The procedure to change the configuration information, and the configuration information and the other information at the same time is as follows:

  1. Stop the HUB monitoring function using "hanetpoll off" command. See "7.7 hanetpoll Command" for the detail.

  2. Deactivate a virtual interface to change using a "stphanet" command. See "7.3 stphanet Command" for the detail.

  3. Change the setup information and common parameter. (For changing monitoring period, the number of monitoring times, recovery monitoring period, cluster switching and link up period, apply changes with "hanetpoll on" command.)
    See "7.7 hanetpoll Command" for the detail.

  4. Deactivate temporarily all virtual interfaces set in NIC switching mode using a "stphanet" command, then reactivate them using a "strhanet" command. See "7.2 strhanet Command" and "7.3 stphanet Command" for the detail.

  5. Starts a function to monitor HUB using a "hanetpoll on" command.
    (For changing monitoring period, the number of monitoring times, recovery monitoring period, cluster switching and link up period, apply changes with "hanetpoll on" command)
    See "7.7 hanetpoll Command" for the detail.

The procedure for enabling a change made on the monitoring information is as follows:

  1. Stop the HUB monitoring function using "hanetpoll off" command. See "7.7 hanetpoll Command" for the detail.

  2. Start the HUB monitoring function to monitor the hubs using "hanetpoll on" command.
    (Changes made to the monitoring period, the number of monitoring times, the monitoring recovery period, the waiting time for a cluster failover, and the waiting time for a link up are reflected when "hanetpoll on" command is executed. For more information, refer to "changing configuration and additional information at the same time".)
    See "7.7 hanetpoll Command" for the detail.

The following lists the information that can be changed for NIC switching mode. No information can be changed besides the information listed below. Delete the concerned definition and add it again.

Note

  • In the cluster environment other than physical IP takeover II, ensure to specify the same IP address configured in "/etc/sysconfig/network-scripts/ifcfg-ethX" when specifying physical IP address by "hanetconfig" command using '-i' or '-e' option. If you specify different physical IP address, it disturbs communication using physical interface because this IP address will overwrite the physical IP address specified with "hanetconfig" command when activating the virtual interface. Do not set any value to IPADDR (IP address) in ifcfg-ethX in the cluster environment with physical IP takeover II.

  • For NIC sharing and tagged VLAN (synchronous switching), in a configuration in which several virtual interfaces share a single physical line, physical interfaces are also inactivated when the last virtual interface is inactivated using the stphanet command.

[Example 1]

The following shows the procedure for changing the monitoring destination of a virtual interface in single system operation.

  1. Check the setting.

    # /opt/FJSVhanet/usr/sbin/hanetconfig print
    [IPv4,Patrol / Virtual NIC]
     Name        Hostname        Mode Physical ipaddr   Interface List
    +-----------+---------------+----+-----------------+---------------------------+
     sha0        192.168.10.100   d   192.168.10.10     eth1,eth2
     sha1        192.168.10.101   d   192.168.10.10     eth1,eth2
    
    # /opt/FJSVhanet/usr/sbin/hanetpoll print
     Polling Status         =  ON        
     interval(idle) =   5( 60) sec        
     time           =   5 times        
     repair_time    =   5 sec        
     link detection =  NO
     FAILOVER Status        = YES
     Name    HUB Poll Hostname
    +-------+--------+------------------------------------------------------+
     sha0       OFF   192.168.10.250,192.168.10.251
     sha1       OFF   192.168.10.250,192.168.10.251
  2. Stop HUB monitoring.

    # /opt/FJSVhanet/usr/sbin/hanetpoll off
  3. Deactivate the target interface. To change the monitoring destinations of sha0 and sha1, deactivate the virtual interfaces of sha0 and sha1.

    # /opt/FJSVhanet/usr/sbin/stphanet -n sha0
    # /opt/FJSVhanet/usr/sbin/stphanet -n sha1
  4. Change the monitoring destination.

    # /opt/FJSVhanet/usr/sbin/hanetpoll modify -n sha0 -p 192.168.10.150,192.168.10.251
    # /opt/FJSVhanet/usr/sbin/hanetpoll modify -n sha1 -p 192.168.10.150,192.168.10.251
  5. Distribute the changes. Because the "IP address of the HUB monitoring destination" was changed in a single system configuration, perform the "deactivate and then activate all the virtual interfaces in NIC switching mode" procedure or "reboot the system" procedure, or "execute the resethanet -s command" procedure according to Procedure 1. The following is an execution example in which the "deactivate and then activate all the virtual interfaces in NIC switching mode" procedure is used.

    # /opt/FJSVhanet/usr/sbin/stphanet 
    # /opt/FJSVhanet/usr/sbin/strhanet
  6. Restart the stopped HUB monitoring. Note that if you performed a reboot or executed the resethanet command, you do not need to perform the following procedure because the monitoring is restarted automatically when GLS reboots.

    # /opt/FJSVhanet/usr/sbin/hanetpoll on

Point

You need to specify the same monitoring destinations for the monitoring destinations of all virtual interfaces sharing the same NIC. Therefore, change all the monitoring destinations at once when changing them.

[Example 2]

The following shows the procedure for changing the virtual IP address of a virtual interface in cluster operation.

  1. Check the setting

    # /opt/FJSVhanet/usr/sbin/hanetconfig print
    [IPv4,Patrol / Virtual NIC]
     Name        Hostname        Mode Physical ipaddr   Interface List
    +-----------+---------------+----+-----------------+---------------------------+
     sha0        192.168.10.100   d   192.168.10.10     eth1,eth2
     sha1        -                p   -                 sha0
    
    # /opt/FJSVhanet/usr/sbin/hanethvrsc print
     ifname     takeover-ipv4    takeover-ipv6     logical ip address list
    +----------+----------------+----------------+--------------------------------+
     sha0:65    192.168.10.100   -                -
  2. Stop the cluster operation and delete the setting for Gls resources from cluster applications.

  3. Stop HUB monitoring and the standby patrol.

    # /opt/FJSVhanet/usr/sbin/hanetpoll off
    # /opt/FJSVhanet/usr/sbin/stpptl -n sha1
  4. Delete the setting for the takeover virtual interface.

    # /opt/FJSVhanet/usr/sbin/hanethvrsc delete -n sha0:65
  5. Change the virtual IP address.

    # /opt/FJSVhanet/usr/sbin/hanetconfig modify -n sha0 -i 192.168.10.101
  6. Set the takeover virtual interface.

    # /opt/FJSVhanet/usr/sbin/hanethvrsc create -n sha0
  7. Create the Gls resource setting on cluster applications.

  8. Distribute the changes. Because the "IP address of a virtual/physical interface" was changed in a cluster configuration, perform a "reboot the system" procedure or "execute the resethanet -s command" procedure according to Procedure 2. The following is an execution example in which the system is rebooted.

    # /sbin/shutdown -r now

[Example 3]

The following shows the procedure for changing the monitoring destination of a virtual interface in cluster operation.

  1. Check the setting.

    # /opt/FJSVhanet/usr/sbin/hanetconfig print
    [IPv4,Patrol / Virtual NIC]
     Name        Hostname        Mode Physical ipaddr   Interface List
    +-----------+---------------+----+-----------------+---------------------------+
     sha0        192.168.10.100   d   192.168.10.10     eth1,eth2
     sha1        -                p   -                 sha0
    
    # /opt/FJSVhanet/usr/sbin/hanetpoll print
     Polling Status         =  ON        
     interval(idle) =   5( 60) sec        
     time           =   5 times        
     repair_time    =   5 sec        
     link detection =  NO
     FAILOVER Status        = YES
     Name    HUB Poll Hostname
    +-------+--------+------------------------------------------------------+
     sha0       OFF   192.168.10.250,192.168.10.251
  2. Stop the cluster operation.

  3. Stop HUB monitoring and the standby patrol.

    # /opt/FJSVhanet/usr/sbin/hanetpoll off
    # /opt/FJSVhanet/usr/sbin/stpptl -n sha1
  4. Change the monitoring destination.

    # /opt/FJSVhanet/usr/sbin/hanetpoll modify -n sha0 -p 192.168.10.150,192.168.10.251
  5. Distribute the changes. Because the "IP address of the HUB monitoring destination" was changed in a cluster configuration, perform a "reboot the system" procedure, or "execute the resethanet -s command" procedure according to Procedure 2. The following is an execution example in which the system is rebooted.

    # /sbin/shutdown -r now

[Example 4]

The following shows the procedure for changing the HUB monitoring interval during single system or cluster operation.

  1. Check the setting.

    # /opt/FJSVhanet/usr/sbin/hanetpoll print
     Polling Status         =  ON        
     interval(idle) =   5( 60) sec        
     time           =   5 times        
     repair_time    =   5 sec        
     link detection =  NO
     FAILOVER Status        = YES
     Name    HUB Poll Hostname
    +-------+--------+------------------------------------------------------+
    s sha0       OFF   192.168.10.250,192.168.10.251
  2. Change the monitoring interval.

    # /opt/FJSVhanet/usr/sbin/hanetpoll on -s 3
  3. To change the monitoring interval, perform a "deactivate and then activate monitoring" procedure or "deactivate and then activate all the virtual interfaces" procedure or "reboot the system" procedure, or "execute the resethanet -s command" procedure. The following is an execution example in which the "deactivate and then activate monitoring" procedure is used.

    # /opt/FJSVhanet/usr/sbin/hanetpoll off
    # /opt/FJSVhanet/usr/sbin/hanetpoll on