Configuring APA in 11.31

[paabhi]

Hi All,

We are planning to configure APA on a new 11.31 server.I read that we can configure either "aggregate" or "LAN MONITOR" method.

My understanding is that In aggregate configuration,both cards will be configured in normal mode and if in case one card fails , the traffic will be taken care by other card.In this method performance and failover is possible,
In "LAN MONITOR" method, one card will always be acting as a standby.

Please could any one tell me the steps for configuring the aggregate method in 11.31

abhi

[Suraj K Sankari]

Hi,
Please check this link below
http://www11.itrc.hp.com/service/james/search.do?docType=Forums&startDateMonth=0&endDateYear=2010&searchString=how+to+configure+apa&prevQueryString=apa+configure&admit=109447626+1275124686756+28353475&dateSearchType=lastXdays&endDateDay=29&dateRange=all&endDateMonth=4&submit1=Search+%C2%BB&resPerPage=20&sortCrit=rank&startDateDay=1&searchCrit=allwords&startDateYear=1999

Suraj

[Abid Iqbal]

Hi,
First show the current configuration:
# nwmgr -g
# nwmgr -g -S apa
Now proceed with the creation of the new failover group.

# nwmgr -a -S apa -c lan900 -A links=0,1 -A mode=aggregate

# nwmgr -s -S apa -A all --saved --from cu

Check
# netstat -ni
# nwmgr -g -S apa

[likid0]

with "aggregate", you make a big fat pipe, with the nic cards you add to the aggregate, so if you add 3 1gig NICS, you get a bandwith of 3 gigs.

with "LAN monitor", you get a primary card and the failover card only are used if the primary fails.

[Tim Nelson]

and some more thoughts.

in an aggregate the switches must support it channel grouping is the common term

scenarios:
two NICs connected to same switch, switch configured for channel group, APA configured on server.
Increased throughput and ha nic failure is the goal

two NICs each in separate switch. switch does not support channel grouping across switches ( typical ). no switch config needed, server is configured for lan failover, active/passive. HA of either a nic or switch failure is the goal.

two NICs each in separate switch, switches are cascaded and configured for channel group. increased bandwidth and ha of either a nic or switch failure is the goal.

hybrid:
two nics cabled to one switch in channel group. two more nics cabled to second switch in a channel group. groupA and groupB are configured in lan monitor. bandwidth and ha are the goal, all the way around.


My config here is the hybrid, a lan monitor of two aggregates/channel groups.

(forgive me if my lan/switch lingo is not exact, hope you simply get the idea )