Mind that: 1) db cannot be safely passed from one thread to another (although probably ok in your example) 2) long running threads have a timeout that depands on the web server configuration
There is an excellent patch from achipa about support for web2y cron jobs. It has not been included yet because I have been slow. Massimo On Jan 4, 7:56 am, Iceberg <iceb...@21cn.com> wrote: > Hi everyone, > > I am writing to share my attempt of "Using threads to do re-occuring > jobs", and seek your advice. > > Basically, I hope to insert some records into a table automatically > every few minutes. This looks same as Voltron did in his post "Re- > occuring jobs". > > http://groups.google.com/group/web2py/browse_thread/thread/1af0256b2b... > > Also, I noticed Achipa's excellent advice about what shall be careful. > http://groups.google.com/group/web2py/msg/e39821790e4b1973 > as well as the discussion in this post "Multiple Threading": > > http://groups.google.com/group/web2py/browse_thread/thread/6252b1e9a7... > Since I plan to use a single web2py (no multi WSGI process), and the > inserting record job should not take long time, so I decide to try > threading anyway. > > I came out with this code snippet, attached at the end of this post. > It is still very rough, but it seems my idea works. The re-occuring > job will be triggered by separate thread, can access same SQLite db > (via main thread, I believe), and other normal requests can be served > during the time. > > I am writing to seek advice. At least I remember Massimo mentioned > "You should not create threads manually by using thread and threading > modules" in his post: > http://groups.google.com/group/web2py/msg/9bf44919adfdbf45 > So, did I miss something important? Say, will this trick cause > infinite threads to be created then eventually halt the main web2py > process? > > My model.py: > db=SQLDB('sqlite://storage.db') > db.define_table('activity',SQLField('whatever','text')) > > All in my controller.py: > > def __stamp(): # do the real job > import datetime, logging > logging.warn('stamp() is called at %s' % datetime.datetime.now()) > rows=db(db.activity.id==1).select(db.activity.ALL) # I can access > SQLite DB in "multithread" world > > def secretEntrance(): # should be protected by only allowing access > from localhost > return __stamp() > > def __trigger(): # Here comes the magic > import urllib,threading > urllib.urlopen('http://%s%s' % ( > request.env.http_host, > URL(r=request,f='secretEntrance')) > ).read() # Use this trick to avoid accessing DB by multi > thread ;-) > threading.Timer(5,__trigger).start() # infinite loop > > def kickoff(): # This is the main entrance. You still need a initial > nudge to trigger all the magic > __trigger() > return 'started' > > def usual_action(): # Just to show that normal requests are served as > usual > return {'foo':'bar'} > > Sincerely, > Iceberg, 2009-Jan-04, 20:54(PM), Sun --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "web2py Web Framework" group. To post to this group, send email to web2py@googlegroups.com To unsubscribe from this group, send email to web2py+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/web2py?hl=en -~----------~----~----~----~------~----~------~--~---
