Have checked the cfaussie list for
a similar problem and Tom's latest problem seemed to be an echo but that
solution did not work!
We need to split our web into a php site and 2
cold fusion sites.
the portal has an application.cfm and uses
/cflogin
Qmis had no application file until I started to
use one to try and solve the problem and does not use cflogin at
all.
The issue is that whichever site you go to first
holds that index.cfm as the one to deliver when you swap to the other
site. it does not matter which order you do it in nor what activity
takes place on the sites. It is only the index.cfm file that gets hooked
- other files within each site deliver okay. The DNS resolution is okay
because hans.htm displays as the right file for each server.
Whatever mess the subdomains get into, the
original subfolders continue to work as they were designed and deliver the
right files.
The 'hold' seems to last for 20 minutes which
argues something to do with the CF server variable timeout. The Cf
server has its own variables (not J2EE) and ap is 2 days; session 20
mins. The web server is IIS (have not checked what version) and its
timeout is 900 secs for the keep alive.
The problem is not session based. That is
once the index file is corrupted for one user on one machine it remains
corrupted for all users until the time out. So it looks like a web
server problem.
I saw a fullasagoog article saying that caching
happened when the cfapplication had loginstorage set to session there were
caching issues but the default for this attribute is cookie anyway and setting
that did not solve anything.
There is no proxy caching in operation on the
network.
Any ideas?
---