B2B 3.6 or 4.0 HA & Clustering

Hi,
Have anyone successfully configured webmethods clustering with either file or datbase repository?

1.Does it require certain type of load balancer if one decides to use one?

2.Does the client code need to program context fetching logic so that it will know how to switch to second B2B server when first B2B server fails?

3.When you use Administrator to add user or package install, how to you make sure the config are done on all the redundant servers?

Thanks in advance for your suggestion.

  1. Q: Does it require certain type of load balancer if one decides to use one?

A: There’s not a requirement to use a certain type of load balancer. Most customers use a h/w load balancer to have a virtual ip address.

  1. Q: Does the client code need to program context fetching logic so that it will know how to switch to second B2B server when first B2B server fails?

A: if you use built in clients (i.e. Context, TContext Java classes) then it’s not necessary. If you use a non webMethods client with a cluster fronted by a load balancer then your client must simply retry against the load balancer. If you use a non-webMethods client with a cluster using native load balancing then your client must have logic to know which server to retry against.

  1. Q:When you use Administrator to add user or package install, how to you make sure the config are done on all the redundant servers?

A: At this time you have to manually configure the servers the same.