> > Can't one use the same scalable/clustered config but make requests
directly
> > - I mean, without this WebDAV client proxy? Please correct me, if I'm
wrong or if I
> > have misunderstood you. Or do you suggest running two application
servers - one for
> > the slide only and another for the custom back-end which calls slide by
means of the
> > WebDAV client - so that both can be load-balanced/clustered/etc.?
> 
> Right. I'm referring to a configuration that we're planning 
> to use with multiple slide servers that are clustered for 
> failover and load-balancing with a single (or multiple) slide 
> WebDAV client-based applications front-ending the servers and 
> a HTTP load-balancer (or tomcat load-balancing)
> in between. In this case if a single slide server goes down, 
> or the load is high, then requests can be farmed out to the 
> rest of the slide cluster for maximum uptime and performance. 
> The slide servers share the same content repository between them.
> With a slide api app that's coresident with the slide server, 
> in the same web container and making calls directly to slide, 
> then I didn't think it was possible for this client app to 
> failover to another slide server (in another app server?) in 
> the event of failure. That is, unless the slide api has been 
> written to use remote calls to find other slide servers on 
> the same machine or another machine? Or does J2EE take care 
> of it somehow??

Well, theoretically Tomcat 5 allows clustering:
http://jakarta.apache.org/tomcat/tomcat-5.0-doc/cluster-howto.html
And a set of tomcats can be load-balanced using various techniques:
http://jakarta.apache.org/tomcat/tomcat-5.0-doc/balancer-howto.html
So you don't need to implement all this explicitly - what you have done in
your WebDAV client-based front-end, right? Do I still miss something? I'm
not an expert in this area, sorry.

> 
> I really don't know enough about the possible scenarios for 
> deploying the slide api and server so I'll ask some questions 
> about your configuration :-)

I understand. Frankly, our current client doesn't need all this
clustering/load-balancing stuff - they have just about 30 users total.
However, our solution could be sold to a couple of other companies who are
considerably bigger so I'm very much interested in your work as well. :-)

> What happens if the app server hosting your slide api client 
> and slide server goes down? Do you have a failover app server 
> to go to in this event?
> Does each failover app server run its own instance of the 
> slide api app and slide server? Do the slide servers share 
> the same content repository?

Right now we just have a failover server which content is synchronized with
the main one's content once per... well... I believe, once per day, but I'm
not sure. :-) It is supposed to be used if the main server crushes.

Yours sincerely,
Andrey.


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to