Monitoring IPMP Information

The examples in this section use the ipmpstat command to enable you to monitor different aspects of IPMP groups on the system. You can observe the status of an IPMP group as a whole or its underlying IP interfaces. You can also verify the configuration of data and test addresses for an IPMP group. Information about failure detection is also obtained by using the ipmpstat command. For more details about the ipmpstat command and its options, see the ipmpstat(1M) man page.

When you use the ipmpstat command, by default, the most meaningful fields that fit in 80 columns are displayed. In the output, all the fields that are specific to the option that you use with the ipmpstat command are displayed, except in the case of the ipmpstat -p syntax.

By default, host names are displayed in the output instead of numeric IP addresses, provided that the host names exist. To list the numeric IP addresses in the output, use the -n option with other options to display specific IPMP group information.

Note - In the following examples, use of the ipmpstat command does not require system administrator privileges, unless stated otherwise.

You use the following options with the ipmpstat command to determine the information to be displayed:

-g displays information about the IPMP groups on the system. See Example 6-9.
-a displays the data addresses that are configured for the IPMP groups. See Example 6-10.
-i displays information about IP interfaces that are related to IPMP configuration. See Example 6-11.
-t displays information about target systems that are used for detecting failure. This option also displays the test addresses that are used by the IPMP group. See Example 6-12.
-p displays information about the probes that are being used for failure detection. See Example 6-13.

The following examples show information about your system's IPMP configuration that you can obtain with the ipmpstat command.

Example 6-9 Obtaining IPMP Group Information

The -g option lists the status of the various IPMP groups on the system, including the status of their underlying interfaces. If probe-based failure detection is enabled for a specific group, the command also includes the failure detection time for that group.

$ ipmpstat -g
GROUP   GROUPNAME   STATE      FDT        INTERFACES
ipmp0   ipmp0       ok         10.00s     net0 net1
acctg1  acctg1      failed     --         [net3 net4]
field2  field2      degraded   20.00s     net2 net5 (net7) [net6]

The output fields provide the following information:

GROUP

Specifies the IPMP interface name. For an anonymous group, this field is empty. For more information about anonymous groups, see the in.mpathd(1M) man page.

GROUPNAME

Specifies the name of the IPMP group. In the case of an anonymous group, this field will be empty.

STATE

Indicates an IPMP group's current status, which can be one of the following:

ok indicates that all underlying interfaces of the IPMP group are usable.
degraded indicates that some of the underlying interfaces in the group are unusable.
failed indicates that all of the group's interfaces are unusable.

FDT

Specifies the failure detection time, if failure detection is enabled. If failure detection is disabled, this field is empty.

INTERFACES

Specifies the underlying interfaces that belong to the IPMP group. In this field, active interfaces are listed first, then inactive interfaces, and finally unusable interfaces. The status of an interface is indicated by the manner in which it is listed:

interface (without parentheses or square brackets) indicates an active interface. Active interfaces are being used by the system to send or receive data traffic.
(interface) (with parentheses) indicates a functioning but inactive interface. The interface is not in use as defined by administrative policy.
[interface] (with square brackets) indicates that the interface is unusable because it has either failed or been taken offline.

Example 6-10 Obtaining IPMP Data Address Information

The -a option displays data addresses and the IPMP group to which each address belongs. The displayed information also includes which addresses are available for use, depending on whether the addresses have been toggled by the ipadm [up-addr/down-addr] command. You can also determine on which inbound or outbound interface an address can be used.

$ ipmpstat -an
ADDRESS         STATE    GROUP      INBOUND     OUTBOUND
192.168.10.10   up       ipmp0         net0    net0 net1
192.168.10.15   up       ipmp0         net1    net0 net1
192.0.0.100     up       acctg1       --          --
192.0.0.101     up       acctg1       --          --
128.0.0.100     up       field2        net2    net2 net7
128.0.0.101     up       field2        net7    net2 net7
128.0.0.102     down     field2       --          --

The output fields provide the following information:

ADDRESS: Specifies the host name or the data address, if the -n option is used with the -a option.
STATE: Indicates whether the address on the IPMP interface is up, and therefore usable, or down, and therefore unusable.
GROUP: Specifies the IPMP interface that hosts a specific data address. Typically, in Oracle Solaris, the name of the IPMP group is the IPMP interface.
INBOUND: Identifies the interface that receives packets for a given address. The field information might change depending on external events. For example, if a data address is down, or if no active IP interfaces remain in the IPMP group, this field is empty. The empty field indicates that the system is not accepting IP packets that are destined for the given address.
OUTBOUND: Identifies the interface that sends packets that are using a given address as a source address. As with the INBOUND field, the OUTBOUND information might also change depending on external events. An empty field indicates that the system is not sending packets with the given source address. The field might be empty either because the address is down or because no active IP interfaces remain in the group.

Example 6-11 Obtaining Information About Underlying IP Interfaces of an IPMP Group

The -i option displays information about an IPMP group's underlying IP interfaces.

$ ipmpstat -i
INTERFACE   ACTIVE   GROUP      FLAGS      LINK       PROBE      STATE
net0        yes      ipmp0      --mb---    up         ok         ok
net1        yes      ipmp0      -------    up         disabled   ok
net3        no       acctg1     -------    unknown    disabled   offline
net4        no       acctg1     is-----    down       unknown    failed
net2        yes      field2     --mb---    unknown    ok         ok
net6        no       field2     -i-----    up         ok         ok
net5        no       filed2     -------    up         failed     failed
net7        yes      field2     --mb---    up         ok         ok

The output fields provide the following information:

INTERFACE

Specifies each underlying interface in each IPMP group.

ACTIVE

Indicates whether the interface is functioning and in use (yes) or not (no).

GROUP

Specifies the IPMP interface name. For anonymous groups, this field is empty. For more information about anonymous groups, see the in.mpathd(1M) man page.

FLAGS

Indicates the status of each underlying interface, which can be one or any combination of the following:

i indicates that the INACTIVE flag is set for the interface. Therefore, the interface is not used to send or receive data traffic.
s indicates that the interface is configured to be a standby interface.
m indicates that the interface is designated by the system to send and receive IPv4 multicast traffic for the IPMP group.
b indicates that the interface is designated by the system to receive broadcast traffic for the IPMP group.
M indicates that the interface is designated by the system to send and receive IPv6 multicast traffic for the IPMP group.
d indicates that the interface is down and therefore unusable.
h indicates that the interface shares a duplicate physical hardware address with another interface and has been taken offline. The h flag indicates that the interface is unusable.

LINK

Indicates the status of link-based failure detection, which is one of the following:

up or down indicates the availability or unavailability of a link.
unknown indicates that the driver does not support notification of whether a link is up or down and therefore does not detect changes in the state of the link.

PROBE

Specifies the state of probe-based failure detection for interfaces that have been configured with a test address, as follows:

ok indicates that the probe is functional and active.
failed indicates that probe-based failure detection has detected that the interface is not working.
unknown indicates that no suitable probe targets could be found, Therefore, probes cannot be sent.
disabled indicates that no IPMP test address is configured on the interface. Therefore, probe-based failure detection is disabled.

STATE

Specifies the overall state of the interface, as follows:

ok indicates that the interface is online and working normally based on the configuration of failure detection methods.
failed indicates that the interface is not working either because the interface's link is down or because the probe detection has determined that the interface cannot send or receive traffic.
offline indicates that the interface is not available for use. Typically, the interface is taken offline under the following circumstances:
- The interface is being tested.
- Dynamic reconfiguration is being performed.
- The interface shares a duplicate hardware address with another interface.
unknown indicates that the IPMP interface's state cannot be determined because no probe targets were found for probe-based failure detection.

Example 6-12 Obtaining IPMP Probe Target Information

The -t option identifies the probe targets that are associated with each IP interface in an IPMP group. The first output is an example of an IPMP configuration that uses test addresses for probe-based failure detection.

$ ipmpstat -nt
INTERFACE   MODE          TESTADDR        TARGETS
net0        routes        192.168.85.30   192.168.85.1 192.168.85.3
net1        disabled      --              --
net3        disabled      --              --
net4        routes        192.1.2.200     192.1.2.1
net2        multicast     128.9.0.200     128.0.0.1 128.0.0.2
net6        multicast     128.9.0.201     128.0.0.2 128.0.0.1
net5        multicast     128.9.0.202     128.0.0.1 128.0.0.2
net7        multicast     128.9.0.203     128.0.0.1 128.0.0.2

The second output is an example of an IPMP configuration that uses transitive probing or probe-based failure detection without test addresses.

$ ipmpstat -nt
INTERFACE   MODE         TESTADDR         TARGETS
net3        transitive   <net1>           <net1> <net2> <net3>
net2        transitive   <net1>           <net1> <net2> <net3>
net1        routes       172.16.30.100    172.16.30.1

The output fields provide the following information:

INTERFACE

Specifies each underlying interface of an IPMP group.

MODE

Specifies the method for obtaining the probe targets.

routes indicates that the system routing table is used to find probe targets.
mcast indicates that multicast ICMP probes are used to find targets.
disabled indicates that probe-based failure detection has been disabled for the interface.
transitive indicates that transitive probing is used for failure detection, as shown in the second example. Note that you cannot implement probe-based failure detection while simultaneously using transitive probes and test addresses. If you do not want to use test addresses, then you must enable transitive probing. If you do not want to use transitive probing, then you must configure test addresses. For an overview, see Probe-Based Failure Detection.

TESTADDR

Specifies the host name or, if the -n option is used with the -t option, the IP address that is assigned to the interface to send and receive probes.

If transitive probing is used, then the interface names refer to the underlying IP interfaces that are not actively used to receive data. The names also indicate that the transitive test probes are being sent with the source address of these specified interfaces. For active underlying IP interfaces that receive data, an IP address that is displayed indicates the source address of outgoing ICMP probes.

Note - If an IP interface is configured with both IPv4 and IPv6 test addresses, the probe target information is displayed separately for each test address.

TARGETS

Lists the current probe targets in a space-separated list. The probe targets are displayed either as host names or IP addresses. If the -n option is used with the -t option, the IP addresses are displayed.

Example 6-13 Observing IPMP Probes

The -p option enables you to observe ongoing probes. When you use this option with the ipmpstat command, information about probe activity on the system is continuously displayed until you terminate the command by pressing Control-C. You must have Primary Administrator privileges to run this command.

The first output is an example of an IPMP configuration that uses test addresses for probe-based failure detection.

# ipmpstat -pn
TIME    INTERFACE   PROBE     NETRTT   RTT      RTTAVG     TARGET
0.11s   net0        589       0.51ms   0.76ms   0.76ms     192.168.85.1
0.17s   net4        612       --       --       --         192.1.2.1
0.25s   net2        602       0.61ms   1.10ms   1.10ms     128.0.0.1
0.26s   net6        602       --       --       --         128.0.0.2
0.25s   net5        601       0.62ms   1.20ms   1.00ms     128.0.0.1
0.26s   net7        603       0.79ms   1.11ms   1.10ms     128.0.0.1
1.66s   net4        613       --       --       --         192.1.2.1
1.70s   net0        603       0.63ms   1.10ms   1.10ms     192.168.85.3
^C

The second output is an example of an IPMP configuration that uses transitive probing or probe-based failure detection without test addresses.

# ipmpstat -pn
TIME    INTERFACE   PROBE      NETRTT   RTT      RTTAVG     TARGET
1.39S   net4        t28        1.05ms   1.06ms   1.15ms     <net1>
1.39s   net1        i29        1.00ms   1.42ms   1.48ms     172.16.30.1
^C

The output fields provide the following information:

TIME: Specifies the time a probe was sent relative to when the ipmpstat command was issued. If a probe was initiated prior to ipmpstat being started, then the time is displayed with a negative value, relative to when the command was issued.
INTERFACE: Specifies the interface on which the probe is sent.
PROBE: Specifies the identifier that represents the probe. If transitive probing is used for failure detection, the identifier is prefixed with either t for transitive probes or i for ICMP probes.
NETRTT: Specifies the total network round-trip time of the probe, measured in milliseconds. NETRTT covers the time between the moment when the IP module sends the probe and the moment the IP module receives the ack packets from the target. If the in.mpathd daemon has determined that the probe is lost, then the field is empty.
RTT: Specifies the total round-trip time for the probe, measured in milliseconds. RTT covers the time between the moment the in.mpathd daemon executes the code to send the probe and the moment the daemon completes processing of the ack packets from the target. If the daemon has determined that the probe is lost, then the field is empty. Spikes that occur in the RTT that are not present in the NETRTT might indicate that the local system is overloaded.
RTTAVG: Specifies the probe's average round-trip time over the interface between the local system and the target. The average round-trip time helps identify slow targets. If data is insufficient to calculate the average, this field is empty.
TARGET: Specifies the host name or, if the -n option is used with the -p option, the target address to which the probe is sent.

Customizing the Output of the `ipmpstat` Command

The ipmpstat command's -o option enables you to customize the output. You use this option with the other previously ipmpstat options to select specific fields to be displayed out of the total fields that the main option normally displays.

For example, the -g option provides the following information:

IPMP group
IPMP group name
Status of the group
Failure detection time
Underlying interfaces of the IPMP group

Suppose that you want to display only the status of the IPMP groups on the system. You would combine the -o and -g options and specify the fields groupname and state, as shown in the following example:

$ ipmpstat -g -o groupname,state
GROUPNAME  STATE
ipmp0      ok
accgt1     failed
field2     degraded

To display all the fields of the ipmpstat command for a specific type of information, include the -o all in the syntax. For example, to list all the fields that provide group information, type ipmpstat -g -o all.

Using the `ipmpstat` Command in Scripts

The -o option is useful when you issue the command from a script or by using a command alias, particularly if you also want to generate machine-parseable output.

To generate machine-parseable information, you combine the -P and -o options with one of the other main ipmpstat options, together with specific fields you want to display. A machine-parseable output differs from normal output in the following ways:

Column headers are omitted.
Fields are separated by colons (:).
Fields with empty values are empty rather than filled with the double dash (--).
When multiple fields are requested, if a field contains a literal colon (:) or backslash (\), these characters can be escaped or excluded by prefixing them with a backslash (\).

To correctly use the ipmpstat -P syntax, observe the following rules:

Use the -o option field with the -P option. Separate multiple option fields with commas.
Never use -o all with the -P option.

Ignoring either one of these rules causes ipmpstat -P to fail.

The following example shows the format of the information when you use the -P option.

$ ipmpstat -P -o -g groupname,fdt,interfaces
ipmp0:10.00s:net0 net1
acctg1::[net3 net4]
field2:20.00s:net2 net7 (net5) [net6]

The group name, failure detection time, and underlying interfaces are group information fields. Thus, you use the -o -g options with the -P option.

The -P option is intended to be used in scripts. The following example shows how the ipmpstat command is issued from a script. The script displays the failure detection time for an IPMP group.

getfdt() {
         ipmpstat -gP -o group,fdt | while IFS=: read group fdt; do
             [[ "$group" = "$1" ]] && { echo "$fdt"; return; }
         done
     }

Skip Navigation Links
Exit Print View
	Managing Oracle Solaris 11.1 Network Performance Oracle Solaris 11.1 Information Library