Link Aggregation on a Single VIOS

The below document explains the need and the process involved in link aggregation. Link aggregation is also called as "Ethernet Bonding". Which ensure continous availability of network service to avoid any failures in the networking service.

The above illustration shows link aggregation by using 3 physical ethernet adapters namely ent0, ent1 and ent2. Link aggregation is made upof primary and secondary adapters.

In this example, the primary adapter is formed my aggregating ent0 and ent1 from the ethernet switch on the left hand side. Another adapter ent2 act as a secondary adapter which is connected to the ethernet switch on the right hand side, to tolerate the failover of the switch on the left hand side.

Configuration:

1) Login into VIOS.

2) $ lsdev -type adapter

name status description

ent0 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

ent1 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

ent2 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

3) Ensure all the attributes of the physical adapters that form the link aggregation must be the same.

# lsdev -dev ent* -attr : Command to view the attributes of an adapter.
# chdev -dev ent* -attr attribute=value : Command to change the attribute of an adapter.

4) Forming Link Aggregation:

$ mkvdev -lnagg ent0,ent1 -attr backup_adapter=ent2

ent3 Available

en3

et3

5) Verify the newly created Link Aggregation.
 ent3 Available EtherChannel / IEEE 802.3ad Link Aggregation

6) View the attributes of the Link Aggregation.

$ lsdev -dev ent3 -attr

attribute value description user_settable

adapter_names ent0,ent1 EtherChannel Adapters True

alt_addr 0x000000000000 Alternate EtherChannel Address True

auto_recovery yes Enable automatic recovery after failover True

backup_adapter ent2 Adapter used when whole channel fails True

hash_mode default Determines how outgoing adapter is chosen True

mode standard EtherChannel mode of operation True

netaddr 0 Address to ping True

noloss_failover yes Enable lossless failover after ping failure True

num_retries 3 Times to retry ping before failing True

retry_time 1 Wait time (in seconds) between pings True

use_alt_addr no Enable Alternate EtherChannel Address True

use_jumbo_frame yes Enable Gigabit Ethernet Jumbo Frames True

$ lsdev -type adapter

name status description

ent0 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

ent1 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

ent2 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

7) Using HMC, create a new virtual ethernet adapter. Ensure the "Access External Network" flag is checked.

$ lsdev -type adapter

name status description

ent0 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

ent1 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

ent2 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

ent3 Available EtherChannel / IEEE 802.3ad Link Aggregation

ent4 Available Virtual I/O Ethernet Adapter (l-lan)

8) Configure Shared Ethernet Adapter using the newly created Link Aggregation (ent3) and Virtual Ethernet Adapter (ent4).

$ mkvdev -sea ent3 -vadapter ent4 -default ent4 -defaultid 2

ent5 Available

en5et5
9) Verify the Shared Ethernet Adapter.

$ lsdev -type adapter

name status description

ent0 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

ent1 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

ent2 Available 2-Port 10/100/1000 Base-TX PCI-X Adapter (1410890

ent3 Available EtherChannel / IEEE 802.3ad Link Aggregation

ent4 Available Virtual I/O Ethernet Adapter (l-lan)

ent5 Available Shared Ethernet Adapter

10) Assign IP over the SEA ent5.

$ mktcpip -hostname <machine name> -interface en5 -inetaddr 10.0.01 -netmask 255.0.0.0

Binding a specific device interrupt to a processor

Every device in AIX has interrupt level. It denotes the way in which the request from the device are submitted to the CPU for execution of task.

Interrupt : It stops the current execution of the processor and executes the interrupted process.

Steps to do binding:

# lsattr -EHl ent0 : Command to view the interrupt level of a device.

# bindprocessor -q : Command to view the processors available in the box.

# lsattr -EHl proc* : Ensure the processors have SMT on. (Simultaneous Multi-threading).

# bindintcpu <interrupt level> <processor no> : Command to bind a interrupt to a specific processor.

# vmstat -i : Command to view the bind interrupt execution.

Do IP_Tracing

IP Tracing utility is handled by the demon called "iptrace".

1) # startsrc -s iptrace -a -s <source machine> -d <destination machine> -p <protocol> -i <interface> <location of the output file>

The above commands monitors/traces all the network related activity happening between the specified source and the destination machine through the specified interface. Finally the output is stored in the desired location.

2) # stopsrc -s iptrace : Command to stop the iptracing demon.

3) # ipreport -srn <output file> : Command to interpet the iptrace output.

TCP/IP Tracing

# iptrace -e -i <interface> <location name> : Command to trace activities on a particular interface and the output is stored in the location.

# ipreport <location of the iptrace output> : Command to view the output of the iptrace command.

Handling Etherchannel Device

# /usr/lib/methods/ethchan_config -f  etherchannel device :  Command to cause maual failover to an etherchannel.

# /usr/lib/methods/ethchan_config -a -b etherchannel backupadapter : Command to add a backup adapter to an etherchannel device.

# /usr/lib/methods/ethchan_config -c etherchannel netadd ping_address : Command to change the address of the control channel.