Auto Port Aggregation setup

[Kenneth Woolard]

I'm having difficulty understanding an internal document on setting up Auto Port Aggregation using 2 GB Lan cards. Here is the extract from the doco :-

###############################################
3.1.5. Auto Port Aggregation Configuration
Next is to edit the Auto Port Agregation configuration file,
/etc/rc.config.d/hp_apaconf
Figure 4. /etc/rc.config.d/hp_apaconf
HP_APA_INTERFACE_NAME[0]=lan900 ## HP default 1st AggPort
HP_APA_LOAD_BALANCE_MODE[0]=LB_MAC ## Load balancing via MAC
HP_APA_HOT_STANDBY[0]=off
HP_APA_GROUP_CAPABILITY[0]=5
HP_APA_KEY[0]= HP_APA_MANUAL_LA[0]=0,1


3.1.6. Network IP Configuration
The final configuration step is to configure the IP connectivity to the Omnipresence network. This is done in similar fashion as you would for normal IP connectivity, except now you will configure a “virtual” interface”
Figure 5. /etc/rc.config.d/netconf
DEFAULT_INTERFACE_MODULES=""
INTERFACE_STATE[0]=up
INTERFACE_NAME[0]=lan900 # as configured in figure 4
IP_ADDRESS[0]=10.97.82.12 # Use YOUR assigned address
SUBNET_MASK[0]=255.255.255.224 # Use YOUR assigned netmask
Note: The number within the [ ] brackets will depend on your hardware configuration, for this example it is 0, as this is the first interface on the system.
3.2. Network Activation
To activate the network connectivity issue the following commands;
# /sbin/init.d/hpapa start
/sbin/init.d/hpapa started
Please be patient. This may take about 40 seconds.
HP_APA_DEFAULT_PORT_MODE = MANUAL
lan900
/usr/sbin/lanadmin -X -l LB_MAC 900
New Load Balancing = 2
/usr/sbin/lanadmin -X -y off 900
New Hot Standby = OFF
/usr/sbin/lanadmin -X -g 900 5 900
New Group Capability = 5
/usr/sbin/lanadmin -X -a 0 1 900
Added ports:0 1 to lan900
/sbin/init.d/hpapa Completed successfully.
##############################################

Questions:-

1. My understanding is this is set just for load balancing an d not failover - Is this true?

2. I have one IP address set for lan 0. Should this same IP address be used for lan 1 i.e. have another entry in netconf
INTERFACE_NAME[1]=lan900

3. Looking at doco there is mention of adding entries to the file 'hp_apaportconf' but in the extract from the doco there is none ?

Basically I'm a little confused and would appreciate clarification.

[VK2COT · Accepted solution]

Hello,

Here are short answers:

1. True. Your setup is designed for
load balancing. Failover is very different.

Assign IP address to only the primary card
in the link aggregate for the failover scenario.

2. No, you assign IP address to the relevant
link aggregate in /etc/rc.config.d/netconf
(lan900).

3. Configure /etc/rc.config.d/hp_apaportconf:
(No entries for these interfaces in the netconf file). For manual change:

HP_APAPORT_INTERFACE_NAME [0]=lan0
HP_APAPORT_GROUP_CAPABILITY [0]=5
HP_APAPORT_PRIORITY [0]=0
HP_APAPORT_CONFIG_MODE [0]=MANUAL

HP_APAPORT_INTERFACE_NAME [1]=lan1
HP_APAPORT_GROUP_CAPABILITY [1]=5
HP_APAPORT_PRIORITY [1]=0
HP_APAPORT_CONFIG_MODE [1]=MANUAL

Example how to put lan0 and lan1 into a
failover group:

HP_APAPORT_INTERFACE_NAME[0]=lan0
HP_APAPORT_CONFIG_MODE[0]=LAN_MONITOR

HP_APAPORT_INTERFACE_NAME[1]=lan1
HP_APAPORT_CONFIG_MODE[1]=LAN_MONITOR

You can configure HP APA so that it will
aggregate ports in any of the following port
aggregation modes:

FEC_AUTO or LACP_AUTO Port Aggregation Modes

MANUAL Port Aggregation Mode

Lot of good references are available at HP.
For example:

http://docs.hp.com/en/J4240-90039/index.html

Cheers,

VK2COT

[Kenneth Woolard]

Thanks for the clarification but I'm still slightly confused on Point 3.
It seems that the way that NIC aggregation (load balancing) is configured means that if one of the NIC fails then all traffic will fall over to the remaining NIC.
To me this seems the same as NIC failover

[VK2COT]

Hello,

Data flows on a failing link are migrated to
the remaining "up" devices in the link
aggregate. When the link returns to the "up"
state, only new data flows will be sent on this
link that has rejoined the link aggregate.

A data flow is defined by the load balancing
algorithm used. In the case of the LB_MAC
algorithm, the data flow is identified by the
low order byte of the destination MAC address.

Cheers,

VK2COT

[Armin Kunaschik]

I don't know how lan_monitor is monitoring the connection. I suspect that it just monitors the link state.
If you setup a balanced configuration it will also work if one link is down. It will not work if you accidentially change the VLAN configuration of one port etc.

The only drawback in the manual mode is, that you have jumping MAC addresses on the switch. lan900 has 1 MAC address and for outging traffic this MAC address is used round robbin on every lan adapter. Some network guys don't like such behavior.

I'd go for the load balancing setup because it includes the failover functionality and adds some kind of (outgoing) port aggregation.

My 2 cents,
Armin