I am having the events logged every 10mins. and the reason for that is OCS client is being disconnected then reconnects back as if nothing has happened.
We have 3 Front-end servers and we are able to identify what's causing it. It seems that it only happens to the last Front-end server started.
We are also using a Cisco LB and if we bypass the ocspool LB VIP (i.e. use any of FE server IP), the issue does not occur.
Where should we start looking at? Is this a LB issue?
Thanks in advance.
Event Type: Information Event Source: Communicator Event Category: None Event ID: 12 Date: 10/27/2008 Time: 8:33:20 AM User: N/A Computer: xxxxx Description: Communicator has enabled event logging. Information about failed calls will be sent to the Windows event log.
For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.
Event Type: Information Event Source: Communicator Event Category: None Event ID: 12 Date: 10/27/2008 Time: 8:23:18 AM User: N/A Computer: xxxxx Description: Communicator has enabled event logging. Information about failed calls will be sent to the Windows event log.
For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.
Try adjusting the timeout variable on the HLB as I've heard of issues like that before where the HLB drops OCS client connections when configured with a default timeout value that is too short.
Did you ever find a resolution for this, we are having the same issue only with CWA and external user's who are being disconnected every 15 mins, we are also using a Cisco LB. Any information would be very helpful, thanks.
Just FYI - I had this problem in a lab environement with 1 Front-End and no load balancer, and I could not figure it out. I never did solve it (even looking through client and server diagnostic logs).
The period of time between disconnect/re-connects varied (sometimes 20 min, sometimes 40 min).
I chalked it up to a lab environment (logged in to a PC with one set of credentials, running Communicator under another against an OCS server in a different forest & domain (with a trust relationship). That theory does not hold up well however because I did not experience this problem for about 7 months and it started happening one day.
