We added 2 more server to our farm (Dell 1855 Blade, 2 INtel Dual Pro 1000 MB cards, that go into a Dell Power Connect 5316. All other blades are using Broadcom dual card. It logs you into the TS but the app does not get published. But if you keep on clicking on the app 2-4 times i can get it to launch. Has anyone seen this?
Error Code Thu Jan 18 12:10:50 2007 - User ssims, Client SALES-4, Address 192.168.11.8 wants to open "TS5 Excel Test" Application, Server NPTS05:3389 is available Thu Jan 18 12:10:51 2007 - User ssims, Client SALES-4, Address 192.168.11.8 connecting directly to 70.xx.xx.34:3389 and will start Application "TS5 Excel Test" Thu Jan 18 12:11:20 2007 - Agent NPTS05, has disconnected Thu Jan 18 12:11:23 2007 - Agent 192.168.1.234 verified successfully Looks like i get connected but the agent gets disconnected before launching the application.Its only happening on 2 of our servers they do have the lastest 4.2.161 V. We have been doing networking monitoring and havent seen anything yet.
2nd Load Balancer issue Another issue related to the agent disconnecting seems to be that the secondary Load Balancer will take over the agent on occasion and not hand it back to the primary load balancer, therefore rendering the Terminal Server unreachable via 2X. I have set the Priority in the registry, but it seems to have no effect on this. I have had to completely disable the secondary load balancer because of this issue. I would really like to be able to run both load balancers at the same time to do some true load balancing, not just as a hot standby.
The agents on the Terminal Servers, as mentioned before are 4.2.161, and the Loadbalancer/Application Servers are both 4.2.161.
An issue has been found in redundancy in version 4.2.161. I sugguest you wait for the next version and give it an other try. It should be out by the end of Jan 2007. Nixu
Suddenly, the terminal server agent on every server with network teaming enabled is exibiting the same issues as the first two blades that had the agent disconnect issue. This is causing major issues, as our clients are now getting error messages stating the server is not a member of the farm, or the windows logon screen comes up, logs them on, then no application launches. No network settings have been changed.
We have now moved the Load Balancer/Publisher to a server that does not have network teaming enabled and the issue seems to have gone away. Maybe this is a timing issue?
Yes, I have seen this. Our storage and exchange clusters are configured without teaming, using a private network for cluster communications. I guess the issue with this program may be similar, though it seemed to work for about a month before problems started to develop. I wonder if the time-out settings could be tweaked via the registry or in the GUI at some point to compensate for this type of setup? Most (not all) of the time, the agent will reconnect in a pretty short period of time.
