Hi, Here's what happened. I'm hosting a web2py application on webfaction (http://interactivepython.org)
This morning, after some maintenance I had tested everything and all was good. Pages were working login/logout was working, database access was working perfectly. Then, because I realized I had stupidly stored my password to the database (Postgresql) out on github, I went to the webfaction dashboard and changed the password for my database. I dutifully made the same change in my configuration file and restarted. Thats when everything came crashing down around me. I was getting the dreaded table already exists error on every request. Changing migrate to false seemed to alleviate the problem, except for two tables where I had not explicitly set migrate to the value in my settings. A little searching through this group is overwhelming in the number of others this seems to effect at various times. So, my question is what happened? My hypothesis is that changing the password outside of the web2py admin is a no no. :-( I haven't figure out how to configure webfaction to allow me admin access. After dropping a couple of the tables that don't have important data in them, I noticed that the prefix on the .table file in the databases directory was different from all the others. So, did changing the password cause the UUID to change? If so, can I recover, and put migrate back to True by renaming all my .table files using the newer prefix? Is there a better way to get things synced up so I could potentially make a schema change? Thanks, Brad
