Using 0.8 with Zope 2.2.4 Chris McDonough wrote:

You're right. Which version of CoreSessionTracking are you using? 0.7 had problems like this that AFAIK 0.8 solves entirely.

----- Original Message ----- From: "Joh Johannsen" <jojo@farm9.com> To: "zope" <zope@zope.org> Sent: Thursday, May 10, 2001 2:54 PM Subject: [Zope] CoreSessionTracking timeout question

...

I've got an Internal Session Data Container timeout set 30 minutes, but occasionally I will start a new session, and all the session data disappears after a couple minutes. The way I understand it, any time I access session data, that 30 minute timer starts again. Is there something I'm doing wrong here?

Thanks,

JJ

_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )

Yes, I noticed this erratic behaviour also and raised the time to 60 minutes and have not had a problem since then, but I would prefer to feel safe setting it back down to 30. On Thu, May 10, 2001 at 04:03:21PM -0400, Chris McDonough wrote:

In that case, I don't know the answer... is anyone else other than Jon having this problem?

------------------------------------------------------ Andres Corrada-Emmanuel Email: andres@corrada.com Internet Programming http://www.mamey.com ------------------------------------------------------

This is troubling. Please keep me informed if weird things start to happen again... before I make the next release I suppose I need to make some more tests. Currently the test suite only tests a session data object with a 1-minute timeout (as a 30 minute timeout would make for a pretty silly test). The 1-minute test seems to work just fine, however. :-( Joh Johannsen wrote:

strange, I changed the timeout to 10 minutes (hoping to find a reproduceable case), and now it has not happened since. I haven't tried 60 minutes yet. 20 and 30 both did not behave very well.

Thanks for the info.

Regards,

JJ

andres@corrada.com wrote:

...

Yes, I noticed this erratic behaviour also and raised the time to 60 minutes and have not had a problem since then, but I would prefer to feel safe setting it back down to 30.

On Thu, May 10, 2001 at 04:03:21PM -0400, Chris McDonough wrote:

...

In that case, I don't know the answer... is anyone else other than Jon having this problem?

------------------------------------------------------ Andres Corrada-Emmanuel Email: andres@corrada.com Internet Programming http://www.mamey.com ------------------------------------------------------

_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )

_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )

Excellent, thanks for the report! I'll check it out... Joh Johannsen wrote:

Good news -- its reproducible, sort of... I've had a brief period where I could reproduce an instant "wipe-out" as soon as a session started. I had a slow-loading page, where it renders minus some graphics -- when it is not completely loaded, I click on a link -- when that link comes up all session data is gone.

If I let the page load completely, session data remains. This makes me think it has more to do with the automatic creation of session data, I do: <dtml-call "REQUEST.set('sessionData', sessionDataManager.getSessionData())"> on every single page -- it seems like sometimes I am just getting a new session when I think I should be getting an existing one.

This way is fairly reproducible, not all the time, but so far around 2/3 of the time (and I've tried it like 20 times). The "losing session data" part is like 2/3 of time, but so far I've never lost it if I let page load completely.

Regards,

JJ

Although we have not bothered to reproduce the timeout in the e-commerce site we are developing, we also are using CoreSession heavily in all of our pages since we have developed a shopping cart system that resides completly in CoreSession (the data is saved to an external database when the customer places the order). On Thu, May 10, 2001 at 08:49:39PM -0700, Joh Johannsen wrote:

Good news -- its reproducible, sort of... I've had a brief period where I could reproduce an instant "wipe-out" as soon as a session started. I had a slow-loading page, where it renders minus some graphics -- when it is not completely loaded, I click on a link -- when that link comes up all session data is gone.

If I let the page load completely, session data remains. This makes me think it has more to do with the automatic creation of session data, I do: <dtml-call "REQUEST.set('sessionData', sessionDataManager.getSessionData())"> on every single page -- it seems like sometimes I am just getting a new session when I think I should be getting an existing one.

This way is fairly reproducible, not all the time, but so far around 2/3 of the time (and I've tried it like 20 times). The "losing session data" part is like 2/3 of time, but so far I've never lost it if I let page load completely.

Regards,

JJ

------------------------------------------------------ Andres Corrada-Emmanuel Email: andres@corrada.com Internet Programming http://www.mamey.com ------------------------------------------------------