I don't know of any real performance problems with Dynamic. It's just theoretically slower because it has to pickle sessions to disk. By the way, it doesn't just pickle sessions to disk when it reaches the limit of 10000, but also after a certain period of inactivity in the session.
Does anyone on the list use the Memory store with the expectation that it pickles sessions to disk on shutdown? If not, maybe you're right, we should just make Memory not save sessions to disk, and document the change very clearly in the release notes. - Geoff > -----Original Message----- > From: Michael Engelhart [mailto:[EMAIL PROTECTED]] > Sent: Tuesday, January 28, 2003 11:19 AM > To: Webware discuss > Subject: Re: [Webware-discuss] how do i prevent a KeyError on > session id > w hen i restart? > > > Yes I saw that patch go through but I didn't like the idea of the > errors just going into the ether and hadn't had time to > response until > now. > > It seems to me that since Dynamic does what Memory does (but with the > added feature of persisting to disk), that we should just > keep the term > Memory and make it transient (i.e, sessions disappear on server > shutdown or timeout). I guess I don't really see the need > for Memory > to persist to disk since I can get that capability by using Dynamic > sessions. If performance is the real reason to use Memory over > Dynamic in a certain application, then maybe we should focus > on fixing > the performance problem associated with Dynamic (which I have yet to > come across myself) instead of moving things around and > changing names, > etc. > I mean the default settings for Dynamic session stores is 10,000 > sessions which I've assumed is an attainable amount before server > meltdown. If you're web application has 10,000 open sessions at any > given time you can probably afford more boxes to cluster your system > around I would guess. > > Personally I like the simplicity of the 3 pronged session > approach. I > just want if fixed :-) > > Mike > > > On Tuesday, January 28, 2003, at 10:55 AM, Ian Bicking wrote: > > > I thought that was added, but I wasn't keeping track. There was a > > session corruption patch which wasn't completely applied -- > it silently > > ignored errors and would have effectively worked like people wanted. > > But it silently ignored errors which wasn't good. > > > > ------------------------------------------------------- > This SF.NET email is sponsored by: > SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! > http://www.vasoftware.com > _______________________________________________ > Webware-discuss mailing list > [EMAIL PROTECTED] > https://lists.sourceforge.net/lists/listinfo/webware-discuss > ------------------------------------------------------- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com _______________________________________________ Webware-discuss mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/webware-discuss
