Servers changing farms from view option !!!!

[DevilDog]

Can anyone duplicate this problem??? I opened a ticket with Citrix but they offered no help unless I allowed them access to our corporate network. No way. This should be easy enough to prove or disprove. Running out of options.


We are currently running MS Terminal Server 4.0 (Build 419) with SP6 and Citrix MetaFrame 1.8 (Build 663) for TSE with SP2 and with/without FR1. All servers are not at the same HotFix level but symptoms are identical.

Problem (short description): Servers are migrating from farm to farm through the View feature of the Published Application Manager.

Problem Description: Server WSR08 is a member of the Production farm. It has published applications. I executed Published Application Manager while residing on server WSR08. The view reflected "[All Servers] [Production]". I chose View/Select Scope from the menu. I insured the option "Within a Citrix Farm" was selected. Under "Choose Server Farm" I selected Staging and clicked OK. At this point the WSR08 server migrated with all published apps to the Staging server farm. As a test, I executed my local desktop PN client which reflected this change. The PN client is v6.01.963.
Also for documentation purposes, we ran "qserver /app /debugnhwatch wsr08" and also against wsr01 which is the Master Browser server. Both of these debug sessions showed that the server, wsr08, changed membership;

06/21 10:25:16 : Neighborhood membership has changed from "Production" to "Staging".. Resetting service.
06/21 10:25:16 : Neighborhood Watch is going down.
06/21 10:25:16 : FindNewServersThread: Exiting
06/21 10:25:16 : Telling WSR02 that I'm not in the farm anymore.
06/21 10:25:16 : I will no longer consider server "WSR02" is a member of my farm.
06/21 10:25:16 : Telling WSR01 that I'm not in the farm anymore.
06/21 10:25:16 : I will no longer consider server "WSR01" is a member of my farm.

etc, etc.......

I was able to duplicate this several times with different servers and different farms. Some servers had Citrix FR1 and some not. All servers were at the same OS level including SP's and same Citrix MF level including SP's.

 

[CitrixEngineer]

I know that running Citrix servers at different levels of Hotfix can produce unpredictable results - but this looks more like a network/routing issue; Are all your Citrix servers on the same subnet? Is the server in question multi-homed? Or is it something simple like you haven't applied the 269214 Hotfix from Microsoft?

I hope this helps

 

[DevilDog]

I have 17 Terminal/MF servers, all on the same subnet and none are multi-homed. MS Hotfix 269214 is installed, but not on all servers. In my environment I am not aware of any of the problems identified in this particular Hotfix. I do not believe that this Hotfix fits my problem, but... I will update all servers with it, just in case. There are other MS and/or Citrix Hotfixes installed but not equal across the board.

What I do not understand from your response is that if it is a network/routing issue than would qserver /app /debugnhwatch pick up the change?

I would love to see someone else with multiple farms try this. -- Devil Dog --

 

[CitrixEngineer]

The reason I suspect the ICA browser is because that is where all farm membership/server/published app/licensing information is held.

I don't wish to teach my grandmother to suck eggs (as the phrase goes), but the browser service simply updates a list held on the Master Browser every time a change is made to a Citrix server. If the information doesn't get to the MB, for whatever reason over a period of time, then the MB will assume that the server no longer holds that information.

I'm sure you've got good reasons, but I would stongly urge you have all your MF servers at an identical build level - including hotfixes. Since you have 17, I assume one is a dedicated Master Browser? If not, then try that as a possible solution.

I hope this helps (At the very least, I've bumped this back to the top so others can make further suggestions )

 

[Highland]

Yeah, sounds like some form of timeout either 2 or from the Master Browser, are all your servers in the same location....???? It could even be that you are experiencing problems with the election process for the master browser, it is possible to set up your servers to be either Master or Potential browsers, I think there are documents on the Citrix Site to detail this....

Might also be worth checking the volume of traffic on your sub-net, if its exceptionally high then this might impact on the browser functionality.....

What about your network hardware...???? could it be a hub/cabling fault....????

 

[DevilDog]

Thanks again. As I said, I will attempt to update all my servers to the same hotfix level during this week. I would still like to hear from someone that has multiple farms and has tried to duplicate this problem. -- Devil Dog --