Network connections on an M2002 appliance
The appliance has a number of network connections.
- 1
- Management Ethernet port mgt0
- 2
- Management Ethernet port mgt1
- 3
- 10 Gb ethernet port, eth0
- 4
- 10 Gb ethernet port, eth1
- 5
- 40 Gb Ethernet module, eth30
- 6
- 40 Gb Ethernet module, eth31
- 7
- 40 Gb Ethernet module, eth32
- 8
- 40 Gb Ethernet module, eth33
- 9
- 1 Gb Ethernet module, eth10
- 10
- 1 Gb Ethernet module, eth11
- 11
- 1 Gb Ethernet module, eth12
- 12
- 1 Gb Ethernet module, eth13
- 13
- 1 Gb Ethernet module, eth17
- 14
- 1 Gb Ethernet module, eth16
- 15
- 1 Gb Ethernet module, eth15
- 16
- 1 Gb Ethernet module, eth14
- 17
- 10 Gb Ethernet module, eth23
- 18
- 10 Gb Ethernet module, eth22
- 19
- 10 Gb Ethernet module, eth21
- 20
- 10 Gb Ethernet module, eth20
You can configure each of the interfaces when you run the installation wizard for the appliance. Alternatively, you can configure them by using the CLI or the IBM® MQ Appliance web UI (see Configuring the appliance).
Aggregate linking and configuration for VLANs is supported for connections. (VLAN is not supported where the connections are used in a high availability or disaster recovery configuration.)
Each 40 Gb Network module provides a total of 40Gb of bandwidth that is shared between the two physical ports on that module. You can create an aggregate interface using one physical connection to each network module (for example, eth30 and eth32) to boost total bandwidth and the availability characteristics of the link.
IPMI LAN connection
You can use an Intelligent Platform Management Interface (IPMI) connection to interface with the Baseboard Management Controller (BMC) on the appliance. The BMC enables you to power up or power down an appliance remotely (provided that power is available in the rack where the appliance is installed). You can also use IPMI commands to perform other various functions, such as printing sensor data and reading event logs. See https://linux.die.net/man/1/ipmitool for details of IPMI commands.
You can use mgt0 or mgt1 for your IPMI connection. You can dedicate the mgt connection to IPMI, or you can share the mgt connection (for example, with administration interfaces). If you share the connection, IPMI requires the allocation of a second IP address in the same VLAN and subnet as the mgt interface.
Because of the nature of IPMI and the level of control that can be obtained over the IBM MQ Appliance that uses this interface, either connect the mgt port to a secure management network, or do not connect it at all.
Administration
You can administer the appliance by using a command line interface (connecting via SSH) or a web user interface. You can restrict these interfaces either to mgt0 or mgt1 to create a restricted management network.
High availability
- eth13 - primary link
- eth17 - secondary link
- eth21 - replication link (advanced users can define a different interface for replication, if required, see Configuring custom HA replication interfaces)
Interface | quirinus | romulus |
---|---|---|
eth13 | 192.168.13.149 | 192.168.13.150 |
eth17 | 192.168.17.149 | 192.168.17.150 |
eth21 | 192.168.21.149 | 192.168.21.150 |
Disaster recovery
If you are implementing a disaster recovery (DR) configuration, note that eth20 is dedicated to replication for DR (advanced users can define a different interface for replication, if required, see Configuring custom DR replication interfaces). The eth20 interface on each appliance in the DR configuration can be in a different subnet, although the subnets should be dedicated to DR replication.
Disaster recovery on high availability systems
If you are implementing DR on an HA system, then eth20 is used for DR replication by both HA appliances, and the recovery appliance (advanced users can define a different interface for replication, if required, see Configuring custom DR replication interfaces).
- The IP addresses used for the two HA appliances should belong to the same dedicated subnet.
- The appliance at the recovery site does not need to belong to the same subnet, but must be able to reach it.
- The two HA appliances can share a floating IP address for eth20 (in addition to each having a static IP address defined for eth20). The floating IP is then used for replication with the recovery appliance by whichever appliance is running the queue manager being replicated (that is, the primary appliance for the queue manager).
- You do not have to physically configure the floating IP. You specify it as an argument when you configure disaster recovery for an HA pair (see Configuring disaster recovery for a high availability queue manager by using the command line). Choose an unallocated IP address on the same subnet as the two static IP addresses.
You can also configure disaster recovery between two pairs of HA appliances. See Example network set up for HA/DR configuration with two HA groups for an example of this configuration.
Naming interfaces
The IBM MQ Appliance enables you to specify a host alias for a specific IP address that is assigned to a network interface. You can use this alias to reference the interface, rather than explicitly using the IP address. Using aliases makes it easier to copy a configuration to another appliance, or migrate between environments, without making extensive changes to accommodate different IP addresses.
data-int; you can use this host alias in IBM MQ commands instead of explicitly referencing the IP address 10.61.121.5. The following example shows the command that is used to create a listener and bind it to the data interface:
define listener(CHA2L) trptype(TCP) control(QMGR) IPADDR(data-int)
If you
then start the listener and display the status, you can see that data-intwas resolved to IP address 10.61.121.5:
start listener(CHA2L)
display lsstatus(CHA2L)
AMQ8631: Display listener status details.
LISTENER(CHA2L) STATUS(RUNNING)
PID(43918) STARTDA(2016-05-04)
STARTTI(09.31.56) DESCR( )
TRPTYPE(TCP) CONTROL(QMGR)
IPADDR(10.61.121.5) PORT(1414)
BACKLOG(100)
For more information about defining host aliases, see Host Alias commands.
Interface | IP address | Comment | Host alias |
---|---|---|---|
eth10 | 10.61.121.5/24 | Used for IBM MQ data | data-int: 10.61.121.5 |
eth11 | 10.161.121.5/25 | Used for monitoring | log-int: 10.161.121.5 |
eth13 | 192.168.121.5/29 | Used for HA Primary | hap-int: 192.168.121.5 |
eth17 | 192.168.122.5/29 | Used for HA Secondary | has-int: 192.168.122.5 |
eth21 | 192.168.123.5/29 | Used for replication | har-int: 192.168.123.5 |
Data connections
You must configure one or more Ethernet connections for the IBM MQ data handled by the appliance. You can use link aggregation to improve the resilience and bandwidth of your data connection.
The M2002 appliance has four 10 Gb connections (eth0, eth1, eth22, eth23) and four 40 Gb connections (eth30, eth31, eth32, eth33) that are not by default used for HA or DR.
Example
The following table shows the network configuration of an example appliance. The appliance is part of an HA group, and also supports disaster recovery for queue managers. IBM MQ data is carried on link aggregated 10 Gb connections, logging data is sent to link aggregated 1 Gb connections. In this example, custom 40 Gb interfaces have been used for HA replication, and DR replication in place of the default eth20 and eth21 replication interfaces. A separate interface (aggregated for availability) is reserved for monitoring data such as writing to syslog targets, or SNMP infrastructure.
Interface | Used for |
---|---|
mgt0 | Web UI and IPMI |
mgt1 | Command line access (SSH) |
eth0 | Not used |
eth1 | Not used |
eth10 | Not used |
eth11 | Not used |
eth12 | Link aggregated 1 Gb interface for monitoring |
eth13 | HA primary group interface |
eth14 | Not used |
eth15 | Not used |
eth16 | Link aggregated 1 Gb interface for monitoring |
eth17 | HA group alternative interface |
eth20 | not used |
eth21 | not used |
eth22 | Link aggregated 10 Gb data interface |
eth23 | Link aggregated 10 Gb data interface |
eth30 | Aggregated with eth32 for custom HA replication |
eth31 | Aggregated with eth33 for custom DR replication |
eth32 | see eth30 |
eth33 | see eth31 |