> Hi, I think you are confused on how synchronization
> works.  The Deployment Manager synchronizes with the
> node agents.  Please see:
> http://www-1.ibm.com/support/docview.wss?uid=swg211084
> 07
>
> You can force a synchronization with the nodes from
> the console or using wsadmin.sh, as long as the nodes
> are managed.  Please see the following for the
> wsadmin commands:
> http://publib.boulder.ibm.com/infocenter/dmndhelp/v6rx
> mx/index.jsp?topic=/com.ibm.websphere.nd.doc/info/ae/a
> e/txml_sync.html
>
> The Application Servers are ready to process requests
> as soon as startup is completed.
>
> Brian




Thank you for your reply! My question is based on the following background.
When a cluster member,say A, is suspended, other cluster members are busy wi
thing handling the requests. So the state kept in A can be significantly dif
ferent with the state of other cluster members. So when A is resumed, reques
ts that belong to some sess
ion can be routed to A if the session owner cluster member fails. In this ca
se, can A still  handle the requests correctly?
Can A get the session information itself from the session backup cluster mem
ber(when using peer-to-peer)? If not, A would start a new session for the re
quest,which would cause errors.

[ Post a follow-up to this message ]