| 5:43 pm on Aug 6, 2005 (gmt 0)|
All these stats have to do with only one page file's records? Not the whole log, right? There were 1490 requests for that pagefile?
| 7:20 pm on Aug 7, 2005 (gmt 0)|
That's correct. 1490 is the page visits for just that page. All the numbers we got from running queries on the database.
| 2:04 pm on Aug 9, 2005 (gmt 0)|
If it were me I would start with the fact that when a cookie is issued by a page it may not actually show up in the hit for that page. I don't have a lot of experience with sites that issue cookies sporadically like this, but I would take that day's logs, throw away lines not having to do with page files, and sort it with Access or Excel into sessions (based on a concatenation of IP and UA field plus time). I'd examine each of those visits that contained your cookied page, and see what's happening in the cookie field. You might find that your cookie is happening but not showing up until the next page view's log line. If it is, that's your answer. If on the other hand it's tending to not show up at all, while the ASPSESSION cookie is, I'd start thinking about your custom cookie's programming.
| 5:23 pm on Aug 9, 2005 (gmt 0)|
The first request for a page that adds a cookie WILL NOT include the cookie. Think of it this way:
A user requests a page. Your server records the request in the log. The response includes a command to save a cookie in with the HTML, etc.
For the second request (usually an image on the page) the request includes the cookie (all requests include the cookie value when it exists). This cookie is dutifully recorded in the log along with all the usual data.
| 5:33 am on Aug 10, 2005 (gmt 0)|
cgrantski and Ben, thanks so much for your inputs.
No, I didn't realize the bit about first request not including the cookie. I think both your inputs combined answers my question to some extent. Testing cookie for just one page probably was not a good idea. We will do it on a few more pages and then collect the results again. I will post the results here again
| 5:36 am on Aug 10, 2005 (gmt 0)|
One more thing I can try. Look for all records that have this cookie page as a referrer. By this time, the records should have cookies. Right? I will do that and post the info here.
| 11:29 am on Aug 10, 2005 (gmt 0)|
| 7:27 am on Aug 11, 2005 (gmt 0)|
Just wanted to update based on what i learned from this thread. I decided to analyse pages that come after the page from where the cookie was inserted.
Page A: the page where we were setting cookie
Page B: pages that were referred by page A
I looked at all records for Page B and these are the results
Records that have the values we explicitly inserted:610
Records that have the session id in the cookie, but not the values we inserted: 237
The good news is that around 72% visits are showing cookies correctly. What we are not able to find is why certain records have asp sessionids showing up in the [cs(cookie)] field in the log file, but not the values we are inserting in the cookie. I tried looking at client ips to see if there is one IP that is skewing the data. That's not the case. They are different IPs.
Hope I'm not confusing anyone here.