Load balance

[IT_Bentley]

We are in the process of configuring failover & load balance on coldfusion
MX7 on win2003 server .

We have two win2003 servers which have been configured with hardware load
balance using NLB.

I installed the coldfusion MX 7 software using multiserver option on both
the servers & the installation is successful on both the servers. Our
webserver is IIS.

When I try to register a remote instance in coldfusion for loadbalance, I am
getting an error - " Network Error".

Has anyone successfully configured cluster /loadbalance on MX7 with the above
configuration. Any suggestions??

Thanks

[tlebel]

Try checking out Brandon Purcells articles on NLB. We followed them for the
most part, and got things up and running with only minor variations.
Maybe start here:
[url]http://www.macromedia.com/devnet/mx/coldfusion/j2ee/articles/balancing_j2ee.html[/url]

Also look at his Blogs on his own site BPurcell.com

Tom Lebel

[IT_Bentley]

I finally was able to create a cluster with two local instances & one remote
instance.

I can get the local instances to failover - if one instance goes down, cold
fusion fails over to the other instance. But If both the local instances crash,
the connections to failover to the remote instance. But I am getting
"Internal error 500 - can not connect to Jrun Server" error

How can I get the remote instance to take over??

Any suggestions.

[Click Here]

Hi there,
I have two FMS 2 server that I need to use as a failover and load balance to
maintain 100% uptime. Does anybody know what would be the best practise doing
that or do you know of material to read about this.
Best regards,
Boris

[JayCharles]

The best approach really depends on the type of applications you're deploying.
If the apps are single user (like a recording app), the easiest way is to use a
hardware load balancer. If you're deploying multiuser apps and you need to be
sure you're routing users to the appropriate server, you might consider
building a "gatekeeper" sort of application that handles routing users.

If you can describe the apps you'll deploy, I can make some suggestions.

[Click Here]

Hi again,
Thank you for your answer.
The setup is two FMS 2 servers running on 2 X 64 bit Intel Cpu, 4GB memory in
each box. Each box fiber connected to San and the boxes has a very large link
to the internet. The plan is to host hundreds of videos in various quality (up
to full screen) and users will then stream them to their computers. One of my
plan was to have the router which connects the boxes to the internet ping them
and see if they are alive and which one give a quicker reply and send the user
to that one. I don?t know if this is best practice or if there are some
recommendation from Adobe at least I have not found them.
Thank you very much for your help.
Boris.

[jegrossman]

I have a similar situation. We're serving video on demand only - no apps - from
FMS2. We want to bring up a second server - no edge servers - and load balance
across the two in round-robin fashion. I've seen info about load balancing with
edge servers but I'm wondering if we can't simply configure the external load
balancer to round-robin, checking port 1935. If it is alive, send request to
that server. If not send to other server. We can deal with admin through each
server's real IP - obviously not through the VIP.

Any reason why this wouldn't work?

[newtoflashcom]

The Best way I know is having a Layer 4-7 switch doing a load balancing with
two or more Edge FMS servers and one Origin FMS servers. The FMS Origin server
will be attached to the Storage where the videos are placed. Every time a user
requests the video, Load balancer will turn to one of the Edge servers. If the
edge server is not having the video, it will request the video from origin and
the same is streamed to the client. On Subsequent requests the Edge don't have
to go to origin or to storage.

But in your scenario if you have only two servers, you can have them behind a
Layer4-7 switch.