Clustered config slightly slower

[t16]

Hi,

Ive got an Alfresco cluster, with 2 x standalone share front ends separated off, and notice that its a little slower than a single box solution with the shares running on the repository.

Is there anything I can do to speed it up regarding caching?

Are there any caching settings that can be adjusted for the share front end on its own?


Thanks

[mrogers]

It will depend upon why the two share instances are slower.   

Are you sure it's the cluster configuration that's slowing things down?   Or could it be that it's due to needing to use a real network rather than say a local loop back or a context switch between applications.

What sort of performance figures are you seeing?  Is it a couple of % or more?

[t16]

Its very marginal… in the order of 100-200ms or so!!

I have a load balancer in front of the share front ends, and a load balancer in front of the two Alfresco repo boxes.

Is there any sort of compression or whatever I could enable on the load balancer in front of the repo cluster?

At the moment I have compression enabled on the balancer in front of the share front ends, but nothing enabled on the balancer in front of the repo cluster.

I take it share will request the content via normal HTTP requests, so perhaps enabling gzip compression for flash and other document content might speed it up?

If I do that, Im unsure if having compression enabled on both load balancers is a waste of time, or whether its necessary to keep the stream compressed all the way to the client?!

Still, I've noticed Alfresco gets faster as more and more stuff is cached in memory. So this is good.