Some problems have been reported with pg8000. Can you reproduce the errors with psycopg2.
Can you email me one of the corrupt tickets? I may be able to read them and see what is going on. On Tuesday, 11 September 2012 09:31:46 UTC-5, lucas wrote: > > oh cr#p, because i was able to definitely trace that cPickle.load error to > the unrecoverable error. makes sense now. > > so how can i narrow this terrible thing down. i am using apache httpd > 2.2.15 on centos 6.3, postgresql 9.1 with the web2py contrib pg8000 driver, > and web2py 2.0.8. i have excluded any unnecessary code as much as i could, > pool_size=7 which changed from my first ticket to this unrecoverable, and i > am not doing any hard commit() or rollback() and such. i do a few > db.update, db.insert, but that is the most manual i get, and again the > exceptions are always when a user is logging in. i have also noted that > the server will kick off a user in mid session and the user has to > re-login. so i am not sure it is with the connection to the db or the > session controller. > > perhaps it is the way i have my /etc/httpd/conf.d/default.conf file > configured for multiple domain names. can someone point me to some example > entries for that file for multiple domains using web2py applications? > > but this has to be fixed, one way or the other because it is sucking wind > like crazy and i have been chasing my tail for 3 weeks now. thanx for the > help and i hope some solution avails itself to bring in stability. > please. lucas > --
