Want to highlight a helpful answer? Upvote!
Did someone help you, or did an answer or User Tip resolve your issue? Upvote by selecting the upvote arrow. Your feedback helps others! Learn more about when to upvote >
Did someone help you, or did an answer or User Tip resolve your issue? Upvote by selecting the upvote arrow. Your feedback helps others! Learn more about when to upvote >
Learn, share, and get recognized
Please increase the session timeout to 15 minutes. I have signed in maybe 8 times today alone. That's not useful when trying to describe a problem or explaining something in detail.
Also, this site is dead slow.
Mac Pro, Mac OS X (10.6.7), Hypercard UG!
I think each user should have a global perfernce with the ability to set their own session timeout limit.
I agree with you that the session time out is too short. Some people have estimated that the session time out is 30 minutes.
I do not know why Apple added the session time out. I assume it was some security thought. I do not have any private information in my account beyond my password. I don't understand all the big concern. I think user have enough understanding of security issues to manage a timeout value.
This is another example of making forum usage more difficult for the average user.
Robert
I do not know why Apple added the session time out. I assume it was some security thought. I do not have any private information in my account beyond my password. I don't understand all the big concern. I think user have enough understanding of security issues to manage a timeout value.
I'm attributing to the trojan making the rounds and, in that case, I'll accept whatever time out they think they need to keep the site malware free.
30 minutes? I can't get in a proper reply without this thing timing out. 30 minutes is way beyond what I am experiencing. I'd say 5 to 10 minutes currently.
In any case, it isn't a forum if people keep getting shut out during reasonable reply times.
I didn't say 30 minutes.
And, I have more than that (it appears) with Firefox (I generally do not use Safari); I've not timed it, but if I'm active, no time out; if I walk away for more than 20 - 30 minutes, I have to sign in again.
Uh, that's how a timeout works. And if you walk away from a session timeout of 10 minutes...for 30 minutes, then obviously you're kicked out. It has nothing to do with the browser.
Bio --
I agree that it has nothing to do with the browser.
It started with the new boards.
None of us really know what the cause is.
We're all just other Mac users, helping each other.
All of us have opinions, but no official statement has been forthcoming, as of yet.
I agree that it has nothing to do with the browser.
You would not think so. For awhile, Firefox 3.x would time out, but TenFourFox would not. Apple cured this "problem" around the time of the invalid 9 cookie problem ( about two weeks ago). Maybe things work differently in different browsers.
You could crank up multiple windows in the same browser. View some randomn forum in a window while you are replying. Every once in awhile come to this window & do something.
I run multiple tabs: command + click on link
some web pages take a long time to finish. Maybe this is why it seems I get more than 15 minutes of time. TenFourFox says that just now one page load complete when I got to typing this paragraph.
Robert
PS. This is a well established procedure for these time out problems. I was doing a job interview once upone a time. Every 15 minutes or so the interviewer would press the return key on his terminal. I asked why he was doing that. He said the system would log him off if they was no input for 30 minutes. I wonder what this was costing the firm?
I design systems like these. Timeouts are a property of a session, which is kept on the server. I know that it's pretty slow, and the design has a lot to be desired, but minutes for a page to load? That's nuts. Half of this is WebObjects, but I see jspa in there as well.
I do know that Apple is growing, so maybe this is a way to find people suitable to work there. I don't know, but they need to make this place slick.
Suggestion 1: When you email me a notification with a link in it, have it log me in instead of me reviewing the thread I already know about, and have to click through 3 more clicks, including the original email.
Suggestion 2: Speed. You're probably operating on Red Hat now and not XServes. Make it work.
Suggestion 3: Session timeouts go longer. See thread above. Must test on this browser called Safari. It's made by Apple. Also use this other obscure browser called FireFox. And Chrome. And Exploder.
Suggestion 4: Very bloated. The user experience also holds the amount of information per vertical inch within a browser. This thing looks like a children's blog. I know you like your white space and columns and bubble conversations, but make it reasonable.
The software that runs the site is Jive SBS, presumably running on a cluster of servers. (The site's content is far too large for a single server to handle it all.) Jive is a commercial product not developed by Apple, & is based on a combination of open source & proprietary technologies. Consult the Jive documentation for info about how the load balancer, the cache server, in-memory caching, etc. interrelate to provide site services.
Note that browser support is provided by the software, that session ID's use a chunk of a single server, & that (at least for the most current 4.5 version) heap memory pressure, not time, determines when cache items are evicted.
It is just a guess on my part, but taking these things into consideration I suspect the session timeout is variable, based on demand, & not something the site administrators can set directly.
So you're saying a load balanced or distributed solution that has no external session management? Hence the problem. They might want to choose a better solution, because it's really defeating the purpose of having a session timeout if people can't use the system properly.
I have no idea what you mean by "external session management." Jive provides load balancing, local & nearby cache management, access to & maintenance of the databases, & all the other services necessary for the cluster configuration. Please consult the documentation if you have questions about how all this works -- it is far too complicated to try to cover in a few paragraphs here.
As for a better solution, apparently there is none. Before 2006 the site ran on software from a different company. At that time the site had to be switched to the then current version of Jive to handle the increasing site load, but that version is no longer supported & has recently also become unable to keep up with the site's ever increasing load.
Lasso. Version 8. Many years ago. Sessions, as we've discussed here, are not being handled properly. I'm not interested in reading up on docs of an obvious problem.
This problem only occurs to me very intermittently. For example I logged in this morning about 7:30 and have been back and forth and away from the forum for as much as 45 minutes or longer and not been "timed out". Sometimes it may happen in 15 minutes if I'm not asking a question or replying or just reading threads. I generally use Firefox 4.0.1, but occasionally Safari. In fact, I had thought the problem had been solved, since it had been a pretty consistant problem for quite a while before this week. I think it is in this past week when it improved for me. I have done nothing different.
laverne's mom
Message was edited by: laverne's mom
As I understand it, Lasso is basically a middleware application server & scripting language. From what I can tell it is not a comparable product to Jive. For instance, it appears to have no built-in tools to distribute the load among the nodes of a cluster or manage distributed caches efficiently so I'm not sure what you mean about it having any inherent ability to do this any better than Jive.
And FWIW, the Jive 4.5 documentation provides insights into the inefficiencies of typical methods of managing distributed caches that designers such as yourself might find interesting even if you are designing one-site-per-server implementations.
As I have said, I think the timeout interval depends on the load on the site; IOW, how may users are actively using it at any particular time.
As I understand it, when a user first accesses the site, Jive creates a session on one of the servers running the site. All interaction with that user session goes through that server, using some of its local resources. Each server has to cache data relevant to that session, including data about additions other user sessions are making at the same time. Since that data could be on a different server, there is a huge amount of server-to-server traffic to make sure the data on each server is current when it is sent to the user.
So I suspect what happens is Jive tries to prevent these servers from being overloaded by dumping inactive sessions when a particular serve's load reaches some predetermined point.
session timeout
