Finally found it. The WSGIProcessGroup qubicbackoffice directive should have been placed in the Directory block. Wonder why it's placed there by modwsgi_deploy seeing how it's apparently not processed when out of place! Anyways, thanks for your help, and I hope by posting this I can help someone someday... Cheers, (A very happy) Yannick
On Jul 8, 1:10 pm, yannick <[email protected]> wrote: > On Jul 8, 12:40 pm, "Diez B. Roggisch" <[email protected]> wrote: > > > > > On Wednesday, July 07, 2010 18:31:27 yannick wrote: > > > Hello all, > > > > We are trying to deploy several TG2 applications behind a single same > > > Apache server, with mod_wsgi. Using modwsgi_deploy worked pretty well, > > > and the point we are at now is the following: > > > - each application works fine on its own behind paster and/or apache > > > w/ mod_wsgi > > > - when accessing one application, then the other always produces the > > > error shown below for the second application, whatever the order is > > > done in. As long as we stay on one given app, there doesn't seem to be > > > a problem. However as soon as we switch app, the problem arises and > > > won't go away until we reload apache. Both apps are very similar in > > > the way they use permissions and sqlite DBs, and present the same > > > error after being loaded after the other. > > > > We didn't try to setup anything using virtualenv, just installed the > > > TG2 apps directly on the system (system-wide). The error seems to > > > indicate there is a conflict of columns in the DB but both apps are > > > configured to use separate sqlite files so I don't see how this could > > > be causing the problem. > > > If you had any insight on what might be causing this very weird issue, > > > I would be very very grateful. > > > IMHO your decision to not use venvs is a mistake, but that's unrelated to > > the > > problem at hand. > > > What I think is the issue (Graham undoubtedly will have more insight into > > this) is that your setup shares one python interpreter, instead of isolating > > these. > > > There is ample documentation about that on the mod_wsgi-site, I think > > "Daemon > > Processes" is the thing to look for. > > > For example, our VHost setup looks like this: > > > WSGIDaemonProcess www processes=10 threads=1 python- > > path=/var/lib/ableton/live/lib/python2.5/site-packages > > WSGIProcessGroup www > > > WSGIScriptAlias "/" "/var/lib/ableton/live/app.wsgi" > > > We have a second one, preview instead of www, which runs without a hitch. > > > Diez > > I have been working for quite a few hours on this, and have tracked it > to the process problem indeed. It appears both my applications are > running in the same process group, despite the instructions I have to > the contrary. I do not understand this at all. To simplify to the max > the problem, here is what I have done. I setup two sites in apache, > and activated the modwsgi_deploy test scripts (test.wsgi), and added a > line to them. They now look like this: > > def application(environ, start_response): > status = '200 OK' > output = 'Hello World!' + environ["mod_wsgi.process_group"] > response_headers = [('Content-type', 'text/plain'), > ('Content-Length', str(len(output)))] > start_response(status, response_headers) > > return [output] > > Now, I linked each application to a separate test.wsgi file as so > (apache sites-enabled files): > > Alias /qubic.backoffice/images /usr/local/turbogears/qubic.backoffice/ > qubic/backoffice/public/images > Alias /qubic.backoffice/css /usr/local/turbogears/qubic.backoffice/ > qubic/backoffice/public/css > Alias /qubic.backoffice/javascript /usr/local/turbogears/ > qubic.backoffice/qubic/backoffice/public/javascript > > #Here is the link to the test file > WSGIScriptAlias /test /usr/local/turbogears/qubic.backoffice/apache/ > test.wsgi > > WSGIDaemonProcess qubicbackoffice threads=10 processes=3 display-name=% > {GROUP} > WSGIProcessGroup qubicbackoffice > WSGIScriptAlias /qubic.backoffice /usr/local/turbogears/ > qubic.backoffice/apache/qubic.backoffice.wsgi > > <Directory /usr/local/turbogears/qubic.backoffice/apache> > WSGIApplicationGroup %{GLOBAL} > Order deny,allow > Allow from all > </Directory> > > I did the exact same thing for the other app, changing the /test alias > to /test2, and all "backoffice" references to "medias" > > I reload apache. Whatever order I load test and test2 in, they are > always appearing in the qubicmedia processgroup. However, when I look > in "ps aux", I have both type of processes running... > I can't wrap my head around where my apache config files are faulty, > but they obviously are! > I really appreciate any and all help you can give > Yannick -- You received this message because you are subscribed to the Google Groups "TurboGears" 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/turbogears?hl=en.
