A failure should be transparent to the client (that's the idea.) So no, there
shouldn't be a need for client changes. However, your server app needs to be
"cluster-aware" which will be your biggest challenge.
"Chris" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> Thanks Mike. We use fault tolerant clustering for other things so hopefully
> there won't be too much to configure. We have lots of client server apps that
> need fault tolerance so our company prefers to handle it all at a higher level
> rather than implement it in each app that needs it. (Hopefully I won't be the
> poor fool who has to set it up!!!)
>
> Also do any modifications need to be made to the clients machines?
>
> "Mike Lowery" <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed)...
>> Yes, clustering would provide this functionality but is difficult to set up.
>> It's also a costly solution for this simple problem. Why not just use a
>> Try/Catch statement in your code so that if the first server is unreachable
>> it tries the second?
>>
>> "Chris" <(E-Mail Removed)> wrote in message
>> news:(E-Mail Removed)...
>>>I have client applications sending messages via TCP to a server application
>>>running on a remote machine. I want to have a second remote machine running
>>>the same server app so that if the first machine is unavailable then clients
>>>can use the second machine. Is there any technology that handles this
>>>seamlessly without the client app having to explicitly change to the second
>>>machine? Does clustering do what I need? Does the technology work for all IP
>>>traffic?
>>>
>>
>>
>
>
|