1 of 1 people found this helpful
is this behavior browser specific? did you check it in other browsers?
or may be you are missing something related to session at LB side
Do you have a load balancer in front of the 2 MT? Do you have Session Persistence (Sticky Bits) set for your MT LB Sessions?...if not, what's happening is that your load balancer is sending you to the other web server, which doesn't have a session setup for you, and it's the other web server that's giving you the problem....you need to have sticky bits enabled.
I have the same problem. I am on ITSM7 Version 7.6.04 SP4 201305030241 Hotfix, servers are in server group, I have two loadbalancers. Please see the architecture below:
For web load balancer source address affinity is used with a timeout of 5460 seconds.
For app load balancer no persistene profile is in use.
In the midtier configuration Session Timeout is set to 90min.
Could it cause the problem? What should I do to avoid this?
Have you checked your sso timeout settings ?
Maximum Session Time [ Default will be 120 mins ]
Maximum Idle Time [ Default will be 30 mins ]
- Log in to SSO Enterprise Console as Administrator
- Click Configuration, Global and then Session.
Users who are infected do not use sso, not all users in the system are infected, just a group of them. I've checked their accouonts on different computers and the problem still exist.
Is there something new regarding this problem. We observe something similar.
I think one good Idea is to check the sticky bit however with that wrongly set you will get timeout almost everytime not really randomly like you wrote.
We didn't change anything exact regarding this issue, just updated the midtier (in some hotfixes some corrections were presented like - this probably should correct such behavior). Actually with the midtier is always good to be on the latest version/sp/hotfix.
Don't forget that there could be also network problems in charge. we had some few users still having this problem. Then it turned out they are sitting in one open space and actually connected to one network element which was guilty in some strange way.
I've noticed that sometimes this is in fact browser specific. Not sure what it is that triggers this in some browsers and not others.