SQLSession works great - ZopeSession is a disaster
Hi, we used SQLSession for some years with success and switched to ZopeSession / CoreSession when we changed to Zope 2.6. At this moment the problems with Sessionmanagement began. I am now at the point thinking about a migration back to SQLSession. My Question now: Does SQLSession reliable work with Zope 2.6 and future releases or is this great / fantastic / best Zope product declared as dead?? regards Harald
On Fri, 06 Aug 2004 10:00:03 +0200, Harald Koschinski <harald.koschinski@friatec.de> wrote:
Hi,
we used SQLSession for some years with success and switched to ZopeSession / CoreSession when we changed to Zope 2.6. At this moment the problems with Sessionmanagement began. I am now at the point thinking about a migration back to SQLSession.
I've also had no joy with the Zope Session machinery. Every now and then when I get cycles to spend on it I try and work on it, but I've had no time at all in the last 6 months or so. (Yes, I've fed the updates back to Chris at the time)
My Question now: Does SQLSession reliable work with Zope 2.6 and future releases or is this great / fantastic / best Zope product declared as dead??
I use it on 2.6 under fairly heavy load, and it works fine. Don't mistake "lack of recent updates" with "dead" - there's been no bugs that required a new release, so I've not done so. I've not yet had reason to try it under 2.7 - upgrading to 2.7 will be a significant effort here, because of all the customised deployment and control software we use (a couple dozen ZEO clients, half a dozen ZEO servers, all controlled by a central script, including deployment and the like). If there are issues with it in 2.7, I'd love to know about it, and would be happy to release a newer version if that's the case. Anthony
My Question now: Does SQLSession reliable work with Zope 2.6 and future releases or is this great / fantastic / best Zope product declared as dead??
I use it on 2.6 under fairly heavy load, and it works fine. Don't mistake "lack of recent updates" with "dead" - there's been no bugs that required a new release, so I've not done so.
That's good to hear - I tried to switch back to SQLSession last weekend :-( but had no luck. SQLSession initialized but then the standardsessionmanagement came between. I tried to delete the sessionfolder on the root - but that is not possible. I tried to delete the sessionproduct in ..lib/python/Product ... - no luck - Zope is not willing to start. Can you please give me a hint how to get it running ? Where is your SQLSession-folder placed? Harald
FWIW, the sessioning machinery was rewritten entirely in Zope 2.7.1, so you may find things better in general if you are willing to use 2.7.1 or 2.7.2. - C On Fri, 2004-08-06 at 04:00, Harald Koschinski wrote:
Hi,
we used SQLSession for some years with success and switched to ZopeSession / CoreSession when we changed to Zope 2.6. At this moment the problems with Sessionmanagement began. I am now at the point thinking about a migration back to SQLSession.
My Question now: Does SQLSession reliable work with Zope 2.6 and future releases or is this great / fantastic / best Zope product declared as dead??
regards Harald
_______________________________________________ Zope maillist - Zope@zope.org http://mail.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope-dev )
--On Freitag, 6. August 2004 5:31 Uhr -0400 Chris McDonough <chrism@plope.com> wrote:
FWIW, the sessioning machinery was rewritten entirely in Zope 2.7.1, so you may find things better in general if you are willing to use 2.7.1 or 2.7.2.
But I have the impression that there are still bugs. Yesterday we had a case where the timeout was set to 300 minutes but some sessions disappeared after 20-30 minutes however this is not reproducable and the behaviour is not bound to high load. Andreas
On Fri, 2004-08-06 at 06:23, Andreas Jung wrote:
--On Freitag, 6. August 2004 5:31 Uhr -0400 Chris McDonough <chrism@plope.com> wrote:
FWIW, the sessioning machinery was rewritten entirely in Zope 2.7.1, so you may find things better in general if you are willing to use 2.7.1 or 2.7.2.
But I have the impression that there are still bugs. Yesterday we had a case where the timeout was set to 300 minutes but some sessions disappeared after 20-30 minutes however this is not reproducable
Of course. If it were any other way, it wouldn't be any fun! ;-)
and the behaviour is not bound to high load.
Well, if you do continue to see it, obviously I'd like to know what's going on. Did you see that I applied a patch suggested by Dieter to CVS/SVN a couple of days ago that might account for the SystemError that folks were seeing? Note that after this error, all bets are off with respect to the state of the BTree data structures, so I wouldn't be suprised if things got a bit out of whack. - C
--On Freitag, 6. August 2004 8:39 Uhr -0400 Chris McDonough <chrism@plope.com> wrote:
Of course. If it were any other way, it wouldn't be any fun! ;-)
and the behaviour is not bound to high load.
Well, if you do continue to see it, obviously I'd like to know what's going on. Did you see that I applied a patch suggested by Dieter to CVS/SVN a couple of days ago that might account for the SystemError that folks were seeing? Note that after this error, all bets are off with respect to the state of the BTree data structures, so I wouldn't be suprised if things got a bit out of whack.
Currently we have switched to Dieters Transience implemenation. Due to my upcoming vacation I can't make any further investigations. Andreas
On Fri, 2004-08-06 at 08:43, Andreas Jung wrote:
Well, if you do continue to see it, obviously I'd like to know what's going on. Did you see that I applied a patch suggested by Dieter to CVS/SVN a couple of days ago that might account for the SystemError that folks were seeing? Note that after this error, all bets are off with respect to the state of the BTree data structures, so I wouldn't be suprised if things got a bit out of whack.
Currently we have switched to Dieters Transience implemenation.
Ah well.
Due to my upcoming vacation I can't make any further investigations.
Sure. Don't worry, I'm not going to make you help. ;-) - C
participants (4)
-
Andreas Jung -
Anthony Baxter -
Chris McDonough -
Harald Koschinski