[reportlab-users] strange problem with reportlab and zope

Tyler Tipton tyler.tipton at gmail.com
Mon Oct 24 16:40:00 EDT 2005


I run a plone/zope website that uses python external methods to call
reportlab for PDF generation. Seemingly everything runs according to the
plan, I have a python external method that gathers SQL data and then uses
reportlab to generate the PDF document, then another python script sets the
HTTP headers and returns the data to the browser.

The problem I am having is that every once in a while after one user has
generated a pdf, all users who try to generate one after that time will get
that users account data instead of their intended data. 9 times out of 10
the correct data comes through but then one set of data seems to get 'stuck'
and then that data gets returned regardless of who is logged in until a
certain amount of reports have been generated and then it goes back to
working fine.

I have of course implemented checks against the account credentials to
ensure that only the currently logged in users data should be returned and
yet somehow the wrong data is still coming through which leads me to believe
it is some kind of caching problem but I have tried everything I know
including setting all the proper headers to disable caching and even
completely disabling the caching proxy and I am totally stumped.

The worst part is that I cannot consistently reproduce this error.. I have
ran numerous tests and closely analyzed all my logfiles and it seems to
happen completely randomly.

If anyone has any ideas as to what could be causing this or any suggestions
as to what I can test to zero in on what is happening here I would greatly
appreciate it.

My server setup is something like this if it helps

internet ->
apache machine with ssl+mod_proxy+mod_rewrite+fastcgi ->
zope fastcgi zserver ->
MS sql server (via mxODBC zope adapter)

Thanks in advance,
Tyler
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://two.pairlist.net/pipermail/reportlab-users/attachments/20051024/cb0330d9/attachment.html


More information about the reportlab-users mailing list