Thank you Professor Di Pierro for your direct reply. Since I am new to the community, I'm not sure how to log a ticket. I'll look into that but if you will bear with me, here is perhaps the additional information you are asking for:
I've deployed Web2PY Version 1.98.2 (2011-08-04 00:47:09) The error is routinely reproducible... on four different web browsers (IE, FireFox, Chrome, Safari). Seems to be associated with starting a new session after an old session times out(?). Refreshing the web browser clears the error, but the error always recurs so is not pretty for end users visiting the site, generating the anonymous, in accessible Web2PY ticket URL which only Admin-privileged folks (i.e. me) can view over an SSH tunnel. I have web2py running behind NGINX via uswgi. This is installed on an Amazon EC2 server using web recipes (e.g. http://stackoverflow.com/questions/5335153/how-to-deply-web2py-using-nginx, https://library.linode.com/web-servers/nginx/python-uwsgi/centos-5 and other helpful sites - I have deployment notes available, if they'd be needed). I don't seem to see this problem at all while running the app on my local machine (using the built in Rocket 1.2.2 server) only when in production deployment on the Amazon cloud. The back end database on the Amazon cloud is a remote MySQL image running at Amazon (RDS - http://aws.amazon.com/rds/). I suspect that mysql connections routinely time out in the cloud if they lay unused for some period of time. Not sure how this might trip up the web2py DAL. You have the Traceback from my previous email (the same one occurs each time...). The Traceback isn't too informative as to exactly what part of my (model?) code is the one triggering the error. Is it the DAL class __init__? I could, I suppose, try to chase this down, but perhaps somebody more intimate with the code can better suggest where to look. Do let me know if any other information other than the above would be of use to troubleshooting the problem. I'll try to read up some more about Amazon RDS (and MySQL) to see if there is anything I can do to tweak the MySQL installation to mitigate timeouts. Meanwhile, perhaps somebody out in the web2py community can comment. Thanks again for your direct reply. Cheers Richard Bruskiewich, PhD Scientific Consultant, International Rice Research Institute (IRRI) On Oct 27, 10:51 am, Massimo Di Pierro <massimo.dipie...@gmail.com> wrote: > Which version? Is the problem reproducible? Can you tell us more? > Perhaps open a ticket > > On Oct 26, 11:24 am, RichardBruskiewich > > > > > > > > <richard.bruskiew...@delphinai.com> wrote: > > I'm a recent fan of web2py, having adopted it for some major > > institutional projects in global agriculture. So far, so good - very > > happy - but I have a recurring curious low level database error which > > I'm not sure how to fix. > > > I've already tried trapping my web2py data model in a try... except > > block, inside a loop to retry the database connection (about 20 times, > > or so I think), but this doesn't seem to help. The database > > connection often works... it is just that when the MySQL connection is > > idle for too long, or when a fresh connection is requested for a new > > user session, something fails... DAL doesn't handle it. > > > Here's the background: I'm running web2py in an NGINX server via uwsgi > > (which generally works fine - configured based on good advice online). > > However, this server runs within an Amazon Web Services (AWS) EC2 > > server and accesses a MySQL instance running non-locally as a AWS > > Relational Database (RDB) server. > > > I'm using a DAL pool_size of 10. > > > Here's the error message: > > > Operational Error: (2013, 'Lost connection to MySQL server during > > query') > > > Traceback (most recent call last): > > File "/opt/web2py/gluon/main.py", line 506, in wsgibase > > BaseAdapter.close_all_instances('commit') > > File "/opt/web2py/gluon/dal.py", line 373, in close_all_instances > > getattr(instance,action)() > > File "/opt/web2py/gluon/dal.py", line 1225, in commit > > return self.connection.commit() > > File "/opt/web2py/gluon/contrib/pymysql/connections.py", line 562, > > in commit > > self.errorhandler(None, exc, value) > > File "/opt/web2py/gluon/contrib/pymysql/connections.py", line 184, > > in defaulterrorhandler > > raise errorclass, errorvalue > > OperationalError: (2013, 'Lost connection to MySQL server during > > query') > > > I hope someone can suggest something here. Do I need to tweak some > > MySQL client parameter, or does someone have some advice about Amazon > > RDB usage practicalities in such a situation? > > > Cheers > > RichardBruskiewich > > Scientific Consultant, > > International Rice Research Institute (IRRI)
