We've just migrated from iPlanet to apache 2.0.40. All our cgi's are
perl 5.8.0. Our cgi's take the users' request, send a string to one
of several servers, take the response, parse it and generate dynamic
html output. This application has worked virtually flawlessly since
1997.
Now that we've migrated to apache we are geting complaints that users
are receiving other users' information.
Example: our pages may be made of of two or more sections. In the
upper section we would have data retrieved from a prior request by the
same person and stored as hidden fields on their screens. The they
would request additional information in which we would fabricate
another string, send it to a host (unix), read the results, parse it
then generate more html.
The users (not all but some) are complaining that what they are seeing
is NOT THEIRS. If the click back then submit again, then they see
their own data.
Now, my question is this: is apache NOT keeping the child process
streams separate? It is possible that apache 2.0.40 is inadvertently
providing the wrong stream to the requesting client?
Anyone every seen this before? Any ideas, recommendations, etc
GREATLY appreciated.
TIA
>> Stay informed about: Can Apache 2.0.40 cause this???