Select appropriate contents for each mode.
When using Fast switching mode, determine the information required for configuration of the mode listed in Table 3.2 Configuration information of Fast switching mode.
Components | ||
|---|---|---|
Virtual interface information (1) | Virtual interface name | |
Virtual IP address or host name | ||
Subnet mask | ||
Physical interface information (1) | Physical interface name | |
IP address or host name | ||
Subnet mask | ||
Physical interface information (2) | Physical interface name | |
IP address or host name | ||
Subnet mask | ||
(Repeat for the number of physical interfaces) | ||
(Repeat for the number of virtual interfaces) | ||
Description of each component is as follows:
Set up the following for the number of virtual interfaces.
Specify a name for a virtual interface, which will be assigned to the physical interface used for redundancy. Specify shaX (X represents a number) of this component using "hanetconfig create" command with -n option.
Specify an IP address or host name to be assigned for the virtual interface. The network portion (IPv4) and a prefix (IPv6) of this IP address must be different from the IP address assigned for the physical interface. When using IPv4, use "hanetconfig create" command with -i option to specify the IP address to be allocated for the virtual interface. When using IPv6, configure these in the /etc/radvd.conf file.
When using an IPv4 address, specify the sub network mask value applied to the virtual IP address. If a subnet is not used, this configuration can be omitted. This component is set by using "hanetmask" command. However, this configuration is not necessary if using an IPv6 address.
Set up the following for the number of physical interfaces used for redundancy.
Specify a name for the physical interface. This component can be set using "hanetconfig create" command with -t option (e.g. eth1, eth2 etc).
If using an IPv4 address, specify an IP address or host name to be assigned for the physical interface. The network portion of this IP address must be different from IP address of other physical and virtual interfaces. To set up this component, create the "/etc/sysconfig/network-scripts/ifcfg-physical interface name" file, and then describe the IP address in the file.
If using an IPv4 address, specify a sub network mask value applied to the physical IP address. If a subnet is not used for allocation, this configuration can be omitted. This configuration is written in the "/etc/sysconfig/network-scripts/ifcfg-physical interface name" file. For RHEL8, set the subnet mask by using PREFIX, not NETMASK. Note that, this configuration is not necessary if using an IPv6 address.
Table 3.3 Configuration information of NIC switching mode shows the information required to configure NIC switching mode:
Components | ||
|---|---|---|
Virtual interface information (1) | Virtual interface name | |
Virtual IP address (or host name) | ||
Subnet mask | ||
Physical interface information (1) | Physical interface name | |
IP address or host name | ||
Physical interface information (2) | Physical interface name | |
Standby interface information | Virtual interface name | |
Automatic switching back mode | ||
Monitored remote system information | Primary Monitored remote system IP address or host name | |
Secondary Monitored remote system IP address or host name | ||
HUB-to-HUB monitoring | ||
(Repeat for the number of virtual interfaces) | ||
Description of each component is as follows:
Set up the following for the number of virtual interfaces.
Name a virtual interface to be configured on a physical interface used for GLS. Specify the name using "hanetconfig create" command with -n option, in "shaX" (where X is a natural number) format.
Specify an IP address or host name allocated to the virtual interface. The network portion (for IPv4) or prefix (for IPv6) of this IP address must be the same IP address assigned to the physical interface. This value is specified using "hanetconfig create" command with -i option.
When using IPv4 address, specify the value of a sub network mask used for the virtual IP address. This configuration can be omitted if not allocating a subnet. Set a subnet mask by using "hanetmask" command. When using IPv6 address, it is not required to configure this value.
Set up the following for the number of physical interfaces for redundancy.
Specify a name of the physical interface. This can be specified using "hanetconfig create" command with -t option. (e.g.eth1, eth2 etc)
Specify an IP address or host name assigned to the physical interface. This IP address must be different from the IP address of the other physical and virtual interfaces. In order to specify an IP address for the physical interface, create "/etc/sysconfig/network-scripts/ifcfg-ethX" file and then assign an IP address in the file.
When using Standby patrol function, set up the following. Skip this process if Standby patrol function is not used.
Specify a name to a virtual interface for standby patrol function. Specify it using "hanetconfig create" command with -n option, in "shaX" (where X is a natural number) format.
Setting up the Standby patrol function enables the automatic switch back function when a transfer path recovers from a failure. Specify "q" to "hanetconfig create" command with -m option for using immediate switch-back after a transfer path recovery, or "p" for using standby interface capability.
Set up the following for the number of virtual interfaces. This configuration cannot be omitted.
Specify an IP address or host name of a HUB to be monitored while primary physical interface is being used. This IP address is assigned using "hanetpoll create" command with -p option. If monitoring by ping command is not used, specify "_none_".
Specify an IP address or host name of a HUB to be monitored while the secondary physical interface is being used. This IP address is specified using "hanetpoll create" command with -p option. This step can be omitted. In such case, the same value as primary remote end IP address or host name is applied. Do not specify if "_none_" is specified to the primary monitored remote system IP address.
Indicate whether the HUB-to-HUB monitoring function should monitor a transfer path between the cascaded HUBs or not, when two HUBs are used:
on: monitor between HUBs,
off: do not monitor between HUBs.
The default value is "off". Specify the value using "hanetpoll create" command with -b option.
Table 3.4 Configuration information of Virtual NIC mode shows the information required to configure Virtual NIC mode.
Components | ||
|---|---|---|
Virtual interface information (1) | Virtual interface name | |
Virtual IP address | ||
Subnet mask | ||
Physical interface information (1) | Physical interface name | |
Physical interface information (2) | Physical interface name | |
Monitored remote system information | Primary Monitored remote system IP address or host name | |
Secondary Monitored remote system IP address or host name | ||
(Repeat for the number of virtual interfaces) | ||
Description of each component is as follows:
Set up the following for the number of virtual interfaces.
Name a virtual interface to be configured on a physical interface used for GLS. Specify the name using "hanetconfig create" command with -n option, in "shaX" (where X is a natural number) format. The interface setting file of the virtual interface is created when the virtual interface is set by the "hanetconfig create" command.
Specify an IP address allocated to the virtual interface. This value is defined in the interface setting file (/etc/sysconfig/network-scripts/ifcfg-shaX) of the virtual interface "shaX".
When setting virtual IP address, specify the value of a sub network mask used for the virtual IP address. This value is defined in the interface setting file (/etc/sysconfig/network-scripts/ifcfg-shaX) of the virtual interface "shaX".
In addition, when setting a cluster takeover IP address with the hanethvrsc command, specify the subnet mask also with the hanetmask command.
Set up the following for the number of physical interfaces for redundancy.
Specify a name of the physical interface. This can be specified using "hanetconfig create" command with -t option. (e.g.eth1, eth2 etc)
Set up the following to activate HUB monitoring of the network monitoring function. Note that HUB monitoring is not performed when this information is omitted.
Specify an IP address or host name of a HUB to be monitored while primary physical interface is being used. This IP address is assigned using "hanetpathmon target" command.
Specify an IP address or host name of a HUB to be monitored while the secondary physical interface is being used. This IP address is specified using "hanetpathmon target" command. This step can be omitted. In such case, the same value as primary remote end IP address or host name is applied.
Table 3.5 Configuration information of GS linkage mode shows the information required to configure GS linkage mode.
Components | ||||
|---|---|---|---|---|
Virtual interface information | Virtual interface name | |||
Virtual IP address or host name | ||||
Subnet mask | ||||
Physical interface information (1) | Physical interface name | |||
IP address or host name | ||||
Subnet mask | ||||
Physical interface information (2) | Physical interface name | |||
IP address or host name | ||||
Subnet mask | ||||
(Repeat for the number of the physical interfaces) | ||||
Virtual gateway information | Virtual gateway IP address | |||
(Repeat for the number of the virtual interfaces) | ||||
Remote node information | Remote node name | |||
Virtual IP information | Virtual IP address or host name | |||
Remote host physical IP address information | IP address or host name (1) | |||
Router IP address or host name (1) | ||||
IP address or host name (2) | ||||
Router IP address or host name (2) | ||||
(Repeat for the number of IP addresses) | ||||
(Repeat for the number of virtual IP) | ||||
(Repeat for the number of remote nodes) | ||||
Description of each component is as follows:
Set up the following for the number of virtual interfaces.
A virtual interface name is specified via "hanetconfig create" command with -n option, in "shaX" (where X is a natural number) format.
Specify an IPv4 address or host name to be assigned to the virtual interface. The network portion of this IP address must be different from the IP address assigned to the physical interface. Virtual IP address or host name is specified via "hanetconfig create" command with -i option.
Specify a sub network mask value applied to the virtual IP address. This configuration can be omitted if not allocating a subnet. Set a subnet mask by using "hanetmask" command. When applying a subnet mask, apply the same mask value to all the virtual and physical IP addresses.
Set up the following for the number of physical interfaces for redundancy.
Specify a name for the physical interface. Physical interface name is specified via "hanetconfig create" command with -t option.
Specify an IP address or host name to be assigned to the physical interface. The network portion of this IP address must be different from the IP address allocated to the other physical and virtual interfaces. To set up this component, create the "/etc/sysconfig/network-scripts/ifcfg-physical interface name" file, and then describe the IP address in the file.
Specify a sub network value applied to the physical IP address. This configuration can be omitted if not allocating a subnet. This configuration is written in the "/etc/sysconfig/network-scripts/ifcfg-physical interface name" file. For RHEL8, set the subnet mask by using PREFIX, not NETMASK. If using a subnet mask, apply the same mask value to all the virtual and physical IP addresses.
Set up the following for the number of virtual interfaces.
Specify the IP address of the remote virtual gateway. The network (subnet) portion of the IP address should be the same as the IP address assigned to the virtual interface. This item creates the "/etc/sysconfig/network-scripts/route-virtual interface name" file and sets the information for the virtual gateway static route in the file. In addition, specify the IP address of the virtual gateway with the -g option of the hanetgw create command.
Configure the following for the number of host nodes.
Specify an arbitrary name (within 16 one-bit characters) to identify the node of remote host. Remote host name is specified via "hanetobserv create" command with -n option.
Set up the following for the number of virtual IPs.
Specify a virtual IP address or host name of the remote host. The virtual IP address or host name is specified via "hanetobserv create" command with -i option. Also, the host name and IP address must be defined in /etc/inet/hosts file.
Specify a physical IP address or host name in the virtual IP of the remote host. List these physical IP addresses separated by ',' (commas). Remote host physical IP address information is specified via "hanetobserv create" command with -t option. The IP address and the host name specified here must be defined in /etc/inet/hosts file as well.
When you use remote network communication with GS via router, specify the IP address or host name of the local system's router in the 'router IP address + remote physical IP address' format according to the remote physical IP address information. The host name and IP address need to be defined in the /etc/hosts file as well. You do not need to set this item if you do not use remote network communication.
Table 3.6 Configuration of redundancy mode shows description of common parameters for each mode. This configuration is not necessary when using the default value.
Contents | Fast switching mode | NIC switching mode | Virtual NIC mode | GS linkage mode | Default |
|---|---|---|---|---|---|
Transfer path monitoring interval | A | N | N | N | 5 sec |
The number of constant monitoring prior to outputting message | A | N | N | N | 0 time |
The number of constant monitoring prior to switching cluster | A | N | N | N | 5 sec |
Switching cluster immediately after starting | A | N | N | N | none |
Outputting message (monitoring the physical interface) | A | N | N | N | none |
Standby patrol monitoring period | N | A | N | N | 15 sec |
The number of constant standby monitoring prior to outputting message | N | A | N | N | 3 times |
Monitoring period | N | A | A | A | 5 sec |
The number of monitoring | N | A | A | A | 5 times |
Recovery monitoring period | N | N | N | A | 5 sec |
Cluster switching | N | A | A | A | Yes |
Link up waiting period | N | A | A | A | 60 sec |
Link status monitoring function | N | A | A | N | Yes |
Hostname resolution function | A | A | N | A | Yes |
Automatic start of monitoring | N | N | A | N | Yes |
The number of recovery monitoring | N | N | A | A | Virtual NIC mode: 2 times |
Automatic fail-back | N | N | A | N | No |
Self-checking function | A | A | A | A | Yes |
[Meaning of the symbols] A: Available, N: Not available
The following are description of each of the content.
Specify the transfer path monitoring interval in seconds. The range of the intervals that can be specified is from 0 to 300 sec. If "0" is specified, it will not monitor the transfer path. Initially, it is set to 5 seconds. The transfer path monitoring interval is set using "hanetparam" command with -w option. This feature is available for Fast switching mode.
Specify the number of times for monitoring before outputting the message (No: 800 or 801) if the message needs to be output as a transfer path failure is detected. The effective range of the numbers which can be specified is from 0 to 100. If "0" is specified, it will not output a message. Initially it is set to 0 (does not output any message). This feature is specified using "hanetparam" command of -m option. Note that this feature is only available for Fast switching mode.
Specify whether or not to switch over the cluster if a failure occurs on a whole transfer path of the virtual interface. The effective range of the numbers is from 0 to 100. it will not switch the cluster. When configuring to switch the cluster, set how many times it repeatedly monitors. The range is from 1 to 100. Initially, it is set to 5, meaning that a cluster failover is triggered after continuously detecting the same failure 5 times. This feature is specified using "hanetparam" command with -i option. This feature is available only for Fast Switching.
Specify whether or not to switch the cluster immediately after the cluster starts up. Configure this if a failure occurs in entire transfer path of the virtual interface before the system starts up. The values which can be specified are either "on" or "off". If "on" is selected, cluster is switched immediately after the userApplication starts up. On the other hand, if "off" is selected, the cluster is not switched even after the userApplication starts up. As an initial value, it is set to "off". This setting is specified using "hanetparam" command with -c option. This is available for Fast switching mode.
Configure whether or not to output a message when the status of the physical interface changes (detecting a failure in transfer path or transfer path recover) in the virtual interface. The values which can be specified are either "on" or "off". If "on" is selected, a message (message number: 990, 991, 992) is output. If "off" is selected, a message is not output. Initially, it is set to "off". This setting is specified via "hanetparam" command with -s option. This is available for Fast switching mode.
Specify the monitoring interval (in seconds) of operational NIC for standby patrol function. The values which can be specified are from 0 to 100. If "0" is specified, it will not run monitoring. Note if the user command execution function (using user command when standby patrol fails or detects recovery) is enabled, do not set the parameter to "0". If the parameter is set to "0", the user command execution function will not work. Initially, the parameter is set to 15 (seconds). This setting is specified via "hanetparam" command with -p option. This configuration is available for NIC switching mode with standby patrol function is enabled.
When a failure is detected in a transfer path using the standby patrol function, a message will be output to inform the failure. In this section, specify how many times to monitor until the message (message number: 875) is output. The values which can be specified are from 0 to 100. If "0" is selected, it stops outputting a message and disables monitoring using the standby patrol function. Note if the user command execution function (using user command when standby patrol fails or detects recovery) is enabled, do not set the parameter to "0". If the parameter is set to "0", the user command execution function will not work. Initially, the parameter is set to 3 (times). This configuration is specified via "hanetparam" command with -o option. This is available in NIC switching mode, which uses the standby patrol function. Using this option, the number of monitoring times doubles immediately after the standby patrol starts.
Specify the monitoring period in seconds. The values which can be specified are from 1 to 300. The default value is 5 (seconds). For Virtual NIC mode, 3 (seconds) is set.
This configuration is specified by the following commands:
NIC switching mode
Specify the value by using the "hanetpoll on" command with the "-s" option.
Virtual NIC mode
Specify the value by using the "hanetpathmon param" command with the "-s" option.
GS linkage mode
Specify the value by using the "hanetobserv param" command with the "-s" option.
This feature is available for NIC Switching mode, Virtual NIC mode, or GS linkage mode.
Specify the number of monitoring times. The values which can be specified are from 1 to 300. The default value is 5 (times). This configuration is specified by the following commands:
NIC switching mode
Specify the value by using the "hanetpoll on" command with the "-c" option.
Virtual NIC mode
Specify the value by using the "hanetpathmon param" command with the "-c" option.
GS linkage mode
Specify the value by using the "hanetobserv param" command with the "-c" option.
This feature is available for NIC switching mode, Virtual NIC mode, or GS linkage mode.
Specify the monitoring period when a failure is detected by communication host monitoring for GS linkage mode. The values which can be specified are from 1 to 300. The default value is 5 (seconds). This configuration is assigned via "hanetobserv param" command with -b option. This feature is available for GS linkage mode.
Specify whether or not to use node switching when a failure occurs to every transfer paths.
yes: Node switching is performed when a failure occurs to a whole transfer paths.
no: No node switching is performed when a failure occurs to a whole transfer path.
The default value is "yes".
This configuration is specified by the following commands:
NIC switching mode
Specify the value by using the "hanetpoll on" command with the "-f" option.
Virtual NIC mode
Specify the value by using the "hanetpathmon param" command with the "-f" option.
GS linkage mode
Specify the value by using the "hanetobserv param" command with the "-f" option.
This feature is available for NIC switching mode, Virtual NIC mode, or GS linkage mode only when operating as a cluster.
Specify the time period (in seconds) until the HUB to links up after monitoring starts. The values which can be specified are from 1 to 300. If this option is not specified, then the default value is used. Initial value is set to 60 (seconds). For Virtual NIC mode, it is set to 45 (seconds). If the value is less than the product of monitoring period and monitoring times (monitoring period X monitoring times), then the value is ignored and ends up using the value of the product of monitoring period and monitoring times.
This configuration is specified by the following commands:
NIC switching mode
Specify the value by using the "hanetpoll on" command with the "-p" option.
Virtual NIC mode
Specify the value by using the "hanetpathmon param" command with the "-p" option.
GS linkage mode
Specify the value by using the "hanetobserv param" command with the "-p" option.
This feature is available for NIC switching mode, Virtual NIC mode, or GS linkage mode.
Specify whether to monitor the link state of the NICs in the virtual interface bundles.
NIC switching mode
The link state is monitored at intervals set by using the -s option of the hanetpoll on command, and GLS immediately performs NIC switching when NIC link down is detected. The default value is "yes". Specify this monitoring with the "-l" option of the "hanetpoll on" command.
Virtual NIC mode
The link status is automatically monitored.
This feature is available for NIC switching mode or Virtual NIC mode.
If you enable this function when the host name, not the IP address, is specified for setting GLS, you can assign the IP address of GLS to NICs based on the host file (/etc/hosts) without depending on the OS setting (/etc/nsswitch.conf). The default value is "yes". This configuration is specified using "-h" option of the "hanetparam" command and enabled in Fast switching mode, NIC switching mode, or GS linkage mode.
Specify whether to start the network monitoring function in conjunction with startup of the virtual interface in Virtual NIC mode.
yes: Starts the network monitoring function in conjunction with startup of the virtual interface.
no: Does not start the network monitoring function in conjunction with startup of the virtual interface.
The default value is "yes". Specify the value using "hanetpathmon param" command with - a option. This value is effective in Virtual NIC mode.
Virtual NIC mode
Specify the number of success counts to go back to the normal monitoring after recovery of a monitoring target is detected in the recovery monitoring by the standby patrol of the network monitoring function. The values which can be specified are from 1 to 300. The default value is 2 (times). (The monitoring target is considered as recovered if the standby patrol succeeds twice.) Specify the value using "hanetpathmon param" command with - r option.
GS linkage mode
Specify the number of retry counts to go back to the normal monitoring after recovery of a monitoring target is detected in the recovery monitoring for the real IP of the communication target. The values which can be specified are from 0 to 300. The default value is 0 (times). (The monitoring target is considered as recovered if the ping monitoring succeeds once and no retry occurs.) Specify the value using "hanetobserv param" command with - r option.
This feature is available for Virtual NIC mode or GS linkage mode.
Specify whether to perform the automatic fail-back when recovery of transfer paths between active NICs and standby NICs is detected by using the standby patrol function in Virtual NIC mode.
yes: Performs the automatic fail-back.
no: Does not perform the automatic fail-back.
The default value is "no". Specify the value using "hanetpathmon param" command with - q option. This value is effective in Virtual NIC mode.
If this function is enabled, the operational state of the GLS is monitored periodically. The default value is "yes". This configuration is specified by "-e" option of "hanetparam" command.
The following describes the upper limit of configuration in each mode.
Upper limit of redundant line control methods
The following table lists the upper limit of configuration items set in the redundant line control methods.
Configuration item | Upper limit |
|---|---|
Total number of virtual interfaces and logical virtual interfaces | 64 |
Upper limit of GS linkage mode
The following table lists the upper limit of configuration items set for communication host monitoring for GS linkage mode.
Configuration item | Upper limit |
|---|---|
Maximum number of virtual IP addresses (Note 1) | 128 |
Maximum number of physical IP addresses | 64 |
Maximum number of nodes in which a single virtual IP address can be transferred (Note 2) | 4 |
Note 1) In the environment where GLS is used in a cluster configuration, you need to configure the following virtual IP addresses as monitoring targets:
Virtual IP address of communication target
Virtual IP address of GLS on the cluster standby node
See
For details on setting of monitoring in a cluster configuration, refer to "3.10.1 Monitoring the remote host".
Note 2) Node can be expanded to 16 by editing the configuration file as follows.
/etc/opt/FJSVhanet/config/ctld.param
# # HA-Net Configuration File # # Each entry is of the form: # # <param> <value> # observ_msg 0 observ_polling_timeout 180 max_node_num 4 <- changed |
Note
Executing the "resethanet -s" command or restarting the operating system is required to reflect the change.
