Have a webserver that was shutdown, then started againg, updated with latest security patches and then rebooted. Now all umbraco sites looses login every 30 secs. It's the same problem on 4.0.3, 4.0.2.1 and 4.0.1 sites.
Problem is not related to number of clicks or any specifik sections or actions, login page reappears whether I just login an wait for 30 secs, or I login an uses different functions in the backend.
The server is running 64bit Windows 2008 Web server edition, IIS7, SQL Server Express 2008, all fully patched. All sites are running in .Net 2.0 integrated mode, with asp.net 3.5 config files. The Windows server is virtualized and I actually don't know what the VM OS is.
All sites run in their own app pool, app pools are set to recycle once every 1740 minutes. App pools run under NetworkService Identity. Idle time-out is set to 20mins.
Any ideas? Could this be related to keepalive.aspx and some security setting in IIS, or is there possibly some setting in IIS7 I'm not aware of?
Jesper, not wanting to scare-monger, but have you done a virus/malware scan recently? I only say this because my Father-in-law's PC recently got a bad virus that screwed up all sorts of browser settings in IE/Chrome (but not Firefox).
If you want me to test logging in, to rule out that it's not just you, then DM (@leekelleher) the details (If you trust me with the password - I'll forget it afterwards).
Remember that most browsers on Windows share the windows (=IE) proxy config. Firefox has seperate proxy settings. I'm not sure what other IE configurations the other browser use, but it could be related to your problem..
Thanks all for your suggestions, problem has now been solved, and I'm going to explain what happened even though it is a bit embarrassing:
As I stated first this server is a virtual server that had been shut down for some time. Turns out that after it had been started again the clock was off with exactly the amount of time that was necessary for the usercontext cookie to expire around 30secs after login, which is 23:59:30 or something like that
To my defense I can only say that I did check the clock on the server, but I was so focused on the time, that a I didn't consider the date (I was probably expecting that any time differences would manifest itself in the time as well as the date *blush*).
What I'm still a bit annoyed about is first that the time synchronization just gave up synchronizing, I would have expected that the difference was equaled out over time, but now is more than 2 days after the server was started again, and I had to manually set the date, and then perform a w32tm /resync from the commandline to get the server properly synchronized.
I'm also a bit confused as to why this virtual machine does not read some kind of time from the hardware the machine is running on, I know that it's running on abstracted hardware, but you would think that this abstracted hardware provided some kind of hardware clock for the operating system.
But anyways - problem solved, and since I'm not dealing with virtual server system administration on a daily basis, I'm gonna leave it at that, and get on with my real job ...
Just a little afterthought: All said, I'm not impressed with Firefox 3.6 that obviously does not respect the expiry times set by the cookies, which enabled FF to keep the login even though the cookie should have expired :-o
Glad your problem is solved ... or at least you know what was causing it!
I actually had the same problem with my VMs (using VMware Workstation though), where the guest VM's clock was running super-fast, and it wouldn't stay in sync with the host's clock. There are various config settings to fix it (can't remember what they are right now), but it's "fixable".
Backend login lost every 30 secs.
Have a webserver that was shutdown, then started againg, updated with latest security patches and then rebooted. Now all umbraco sites looses login every 30 secs. It's the same problem on 4.0.3, 4.0.2.1 and 4.0.1 sites.
Problem is not related to number of clicks or any specifik sections or actions, login page reappears whether I just login an wait for 30 secs, or I login an uses different functions in the backend.
The server is running 64bit Windows 2008 Web server edition, IIS7, SQL Server Express 2008, all fully patched. All sites are running in .Net 2.0 integrated mode, with asp.net 3.5 config files. The Windows server is virtualized and I actually don't know what the VM OS is.
All sites run in their own app pool, app pools are set to recycle once every 1740 minutes. App pools run under NetworkService Identity. Idle time-out is set to 20mins.
Any ideas? Could this be related to keepalive.aspx and some security setting in IIS, or is there possibly some setting in IIS7 I'm not aware of?
Regards
Jesper Hauge
Switched on logging of all app pool recycle events, the app pool is not recycling.
Help really appreciated on this one
.Hauge
In fact I set up a performance monitor to monitor application restarts - there is no application restart happening when the login times out.
.Hauge
Hi Jesper, you've checked everything that most of us would suggest.
Taking a long shot, have you tried this with a different web-browser? Could it be a cookie issue?
Try logging in, checking your cookie - then once you get logged out, see what value the cookie has.
Good luck, Lee.
Hmmm I would have guessed at the AppPool recycle settings... But if you have double checked those, then its a bit odd to say the least
Lee K. is on track! Problem is only present in IE8 and Chrome4, not in FF 3.6
Looking into cookie right now.
.Hauge
Jesper, not wanting to scare-monger, but have you done a virus/malware scan recently? I only say this because my Father-in-law's PC recently got a bad virus that screwed up all sorts of browser settings in IE/Chrome (but not Firefox).
If you want me to test logging in, to rule out that it's not just you, then DM (@leekelleher) the details (If you trust me with the password - I'll forget it afterwards).
I'm more than happy to try as well :) Lot of love going round these forums hey... lol
Remember that most browsers on Windows share the windows (=IE) proxy config. Firefox has seperate proxy settings. I'm not sure what other IE configurations the other browser use, but it could be related to your problem..
Thanks all for your suggestions, problem has now been solved, and I'm going to explain what happened even though it is a bit embarrassing:
As I stated first this server is a virtual server that had been shut down for some time. Turns out that after it had been started again the clock was off with exactly the amount of time that was necessary for the usercontext cookie to expire around 30secs after login, which is 23:59:30 or something like that
To my defense I can only say that I did check the clock on the server, but I was so focused on the time, that a I didn't consider the date (I was probably expecting that any time differences would manifest itself in the time as well as the date *blush*).
What I'm still a bit annoyed about is first that the time synchronization just gave up synchronizing, I would have expected that the difference was equaled out over time, but now is more than 2 days after the server was started again, and I had to manually set the date, and then perform a w32tm /resync from the commandline to get the server properly synchronized.
I'm also a bit confused as to why this virtual machine does not read some kind of time from the hardware the machine is running on, I know that it's running on abstracted hardware, but you would think that this abstracted hardware provided some kind of hardware clock for the operating system.
But anyways - problem solved, and since I'm not dealing with virtual server system administration on a daily basis, I'm gonna leave it at that, and get on with my real job ...
Thanks again
.Hauge
Just a little afterthought: All said, I'm not impressed with Firefox 3.6 that obviously does not respect the expiry times set by the cookies, which enabled FF to keep the login even though the cookie should have expired :-o
.Hauge
Hi Jesper,
Glad your problem is solved ... or at least you know what was causing it!
I actually had the same problem with my VMs (using VMware Workstation though), where the guest VM's clock was running super-fast, and it wouldn't stay in sync with the host's clock. There are various config settings to fix it (can't remember what they are right now), but it's "fixable".
Cheers, Lee.
is working on a reply...
This forum is in read-only mode while we transition to the new forum.
You can continue this topic on the new forum by tapping the "Continue discussion" link below.