@Dan - In general you are correct, this is a ColdFusion deployment issue.
What I was curious about is to see if other people that deploy MG
applications came-up with an elegant solution to this problem. Specifically
because the MG apps cache the controller code / MG application code by
default so you are forced to deal with this issue when deploying updates to
the application in MG in a cluster. But in general, you would be forced to
do this with anything that caches objects in the application scope, etc.

@Chris - Each instance does have an internal webserver port. However, when I
tried to access the site with this port it simply returns to the ColdFusion
root. For instance www.sitename.com:8301, www.sitename.com:8302,
www.sitename.com:8303, etc all just return the CFIDE directory and not the
site itself. This would be the ideal option, but I just can't figure out how
to point the server to that specific port.

@Allen - Pulling each instance out, verifying the result and then adding it
back to the cluster has been the recommendation from another CF engineer I
talked too as well. But from what we could determine this remained a manual
process. We are trying to reduce as much manual overhead as possible. We
have a integration machine that matches production 100%, so in theory all of
our functional / unit test run against the build before it's launched. We're
trying to hook the last point of the process into the automated deployment
which would deploy the update, initialize the application and then do a
verification on the server.

The only other alternative we've come-up with here is to make this automated
would be to write a script on the server that slowly cycles each instance,
allow the users from that instance to fail-over and then continue down the
process. Overall, we're just trying to find a solution that doesn't cause
the server to have to do a query or file read on EVERY request just to see
if it should restart. Also, this would be problematic as if we deployed the
update and all instances reloaded at the same time it could bring the
cluster down as they are all refreshing at the same time.

However, since this discussion is slightly off topic I'll take this
discussion off-list for now, and post back the results once we find a good
solution. If anyone has another forum they would suggest I take the
discussion to or would like to continue it with me off-list please contact
me directly.

Thanks for your help,
Roy

-- 
Model-Glue Sites:
Home Page: http://www.model-glue.com
Documentation: http://docs.model-glue.com
Bug Tracker: http://bugs.model-glue.com
Blog: http://www.model-glue.com/blog

You received this message because you are subscribed to the Google
Groups "model-glue" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/model-glue?hl=en

Reply via email to