IMServer:mod wsgi: Difference between revisions

From KitwarePublic
Jump to navigationJump to search
No edit summary
Line 17: Line 17:


* Any wsgi script must be able to maintain persistent connection with mongo to boost speed
* Any wsgi script must be able to maintain persistent connection with mongo to boost speed
* Script to test persistent connections.
 
== Testing persistent interpreter ==


<pre lang="python">
<pre lang="python">


sys.path.append("/var/www-dj/")
sys.path.append("/var/www-dj/")
conn = pymongo.Connection()
 
db = conn["clay1"]
try:
try:
import test6  
import test6  
Line 46: Line 45:
return [str1]
return [str1]
</source>
</source>
The test6.py in the python path contains a single line
<pre>
x = 199
</pre>
This takes advantage of the singleton pattern of python modules.
If test6.py is newly loaded in the run time the value of x will be initialized to 199 every time. And hence the output will always be 200.
If the module fails to load (due to incorrect path etc.) then the value will be 178. And if the module is just loaded once due to persistent interpreter for multiple calls, the value will be initialized to 199 during loading, and will continuously increase in subsequent calls.

Revision as of 16:02, 19 January 2011

wsgi interface in Apache

  • Recompile with the matching python version
  • I had issues with python 2.4.* but worked well with python 2.6.*

When wsgi interface uses modules from the local directory, the script should first append python path as follows otherwise the script which runs without errors in debug mode will not be able to run when called from wsgi interface.

<source lang="python">

sys.path.append('/var/www-dj/') from mongo import m

</source>


Development

  • Any wsgi script must be able to maintain persistent connection with mongo to boost speed

Testing persistent interpreter


sys.path.append("/var/www-dj/")

try:
	import test6 
except:
	class sm():
		def __init__(self):
			pass
			self.x = 177;

	test6 = sm()

def application(environ, start_response):
    	""" The WSGI test application """
    	# emit status / headers
	status = "200 OK"
    	str1 = "New value of X is %d"%(test6.x)
	response_headers = [('Content-type', 'text/plain'),
        	            ('Content-Length', str(len(str1)))]
    	start_response(status, response_headers)
	test6.x = test6.x+1
	return [str1]
</source>

The test6.py in the python path contains a single line 

<pre>
x = 199

This takes advantage of the singleton pattern of python modules.

If test6.py is newly loaded in the run time the value of x will be initialized to 199 every time. And hence the output will always be 200. If the module fails to load (due to incorrect path etc.) then the value will be 178. And if the module is just loaded once due to persistent interpreter for multiple calls, the value will be initialized to 199 during loading, and will continuously increase in subsequent calls.