On Nov 15, 2008, at 10:33 PM, Randy Syring wrote:
> Thank you so much for your response, I am extremely grateful.
> However, I am still getting exceptions thrown from SQLite for sharing
> connections across threads.
Unfortuantely I cant reproduce your error now, whereas it was quite
frequent earlier. Attached is a full test script. If the error
you're getting is only with the ISAPI library you're using, that may
be part of the problem.
>
> Are you sure this isn't a problem with pulling the wrong connection
> from the pool?
> I was able to use a NullPool successfully to alleviate my problems.
> However, it still doesn't seem right that I can't get singleton thread
> pooling to work. Am I still doing something wrong?
The SingletonThreadPool in SQLA 0.5 uses Python's threading.local() to
maintain the connection and is extremely simple. SQLA doesn't spawn
any threads and only works within the thread from which is was
invoked, with the huge asterisk that Python's gc.collect() can invoke
cleanup operations in an asynchronous thread, but even in that case
the SingletonThreadPool never associates the connection with the
threading.local() a second time. Just read the source and try
experimenting with it.
If you do test with too many threads, the Python interpreter itself
may be failing in some way. If I run ab against this script with more
than 20 or 30 threads I get a broken socket almost immediately.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"sqlalchemy" 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/sqlalchemy?hl=en
-~----------~----~----~----~------~----~------~--~---
### sqlalchemyapp.py
import logging
import thread
import time
from sqlalchemy.orm import sessionmaker, scoped_session
from sqlalchemy import create_engine
from sqlalchemy import Table, Column, Integer, String, MetaData, ForeignKey, Unicode
from sqlalchemy.orm import mapper
logging.basicConfig()
logging.getLogger('sqlalchemy.pool').setLevel(logging.INFO)
logging.getLogger('').setLevel(logging.INFO)
engine = create_engine('sqlite:///test.sqlite')
metadata = MetaData()
postings_table = Table('employers_postings', metadata,
Column('id', Integer, primary_key=True),
Column('title', Unicode(150)),
Column('education', Unicode(150))
)
class Posting(object):
def __init__(self,title, education):
self.title = title
self.education = education
def __repr__(self):
return '[Posting "%s" : %d]' % (self.title, self.id)
mapper(Posting, postings_table)
Session = scoped_session(sessionmaker(bind=engine))
## WSGI application, called on each request
def sqlapp(environ, start_response):
# this helps to make IIS use more than one thread
time.sleep(0.5)
sess = Session()
exclogged = False
try:
logging.info('(%d) start response' % thread.get_ident())
if environ['PATH_INFO'] == '/add':
p = Posting(u'test', u'posting')
sess.add(p)
sess.commit()
retval = 'posting added'
elif environ['PATH_INFO'] == '/setup':
metadata.create_all(bind=engine)
retval = 'tables created'
else:
try:
postings = sess.query(Posting).all()
except Exception, e:
logging.info('(%d) query exception: %s' % (thread.get_ident(), str(e)))
exclogged=True
retval = 'Postings: \n\n%s' % postings
start_response('200 OK', [('content-type', 'text/html')])
return [retval]
except Exception, e:
if not exclogged:
logging.info('(%d) exception: %s' % (thread.get_ident(), str(e)))
finally:
try:
Session.remove()
except Exception, e:
logging.info('(%d) finally exception: %s' % (thread.get_ident(), str(e)))
logging.info('(%d) end response' % thread.get_ident())
from wsgiutils import wsgiServer
server = wsgiServer.WSGIServer (('localhost', 8000), {"/":sqlapp})
server.serve_forever()
