On Jan 21, 2011, at 11:13 PM, habumaster wrote: > > Going the servlet route.. Your app is running on a main server, not a virtual > one I am guessing.
I've run this kind of setup successfully on a single machine (OSX), on multiple physical servers (Centos5), and on multiple OpenVZ guests (Centos5 host & guests). > Mine is virtual and with the apache config you spec'd, I understand it was a > starting point. > I have ActiveMQ 5.4.2 running on the same server. Jetty.xml is in the > activemq.xml file, so Ajax servlet should be there. Using the > amq_prototype_adapter.js. > > I have a Ruby app on :8080 on that same server that serves up a page. > The page calls amq.init with "\amq" onload. Using the ProxyPass etc in > Apache with web logs set to debug. > As my GET goes out to \amq, I see in the apache logs the proxy getting setup > to from the web client to localhost:8161 then back from 8161 to /amq. There > is a body send, body end and a release. then the dame thing all over again > in the apache error_log. I can see the client cycling like mad with multiple > GETs to /amq. The Firebug log fills up in no time. > Is this something in particular with the amq.js? > > Did you start at this point as well when you brought your app up with this > setup? > Trying to figure out where to go from here. > Having a little trouble following this problem report. Some more specifics would help a lot, I think. 1. Is your Ruby app Rails or something else? Running via passenger, mongrel, or other? 2. What URL do you visit to view your test web page? 3. Post the <script> tags you have in that page, and the JS code you're using to configure amq.js. 4. Post the proxy configuration you've got in Apache. 5. Post a few examples of the Apache errors you see. 6. Post a few examples of whatever is filling the Firebug log. It would be great to see a GET for polling, and the response which comes back from ActiveMQ. Hopefully something in there will help point out what the problem is. alex
