Ok, here is the scenario. I build a test-environment these days ... as part of a bigger implementation later on. Trying to figure out all there is to figure out about the citrix web-interface.
Short description of my test-environment:
2 citrix metaframe xpa fr3 servers
1 ms sql db-server (which holds the data store db)
1 web server with web interface 2.0 installed
What I want:
Load balancing & the ability to have some degree of fault tolerance.
The load balancing stuff looks pretty straight-forward (after trial and error). But there is still some mysteries about this web-interface product.
For instance, when I install the webinterface I had it point to my metaframeserver1 machine. Afterwards I went into the wiadmin interface and added metaframeserver2 as the second server to use (also as the second machine in the list order).
I do a test logon against both machines (not through the web-interface but through the ica-client) just to check that the servers hangs with me. Then I publish an application (notepad) and do a test logon against metaframeserver1 from the web interface (I guess it is the machine that answers, cause it is the first in the server-list).
Then, I power off the metaframe1 server and tries to log on once more. Nothing happens (should not server2 be the point of authentication right now?). I go to the webinterface admin console (wiadmin) and change the order of my servers so server2 is the first one and server1 is the second one. Try another log-in (same result).
Then I boot up metaframeserver1 again ... I even change back the server order so metaframeserver1 is first in the list, but it still fails to log on.
My impression is that the web interface seems a bit moody (be nice to me, or else I refuse to cooperate
) ). The strange part is that the failed logons do not produce any error-statement in the web-interface. It just pops back to the log-on screen with empty user, password and domain fields.
I should be able to reach my farm even if one of my servers did not respond (for instance due to a power failure) right? And as long as I put the data store on a db-server (independent of the metaframe servers ... they should reach all the stuff they need without needing anything from each other?
Can anyone fill me in on this one?
Short description of my test-environment:
2 citrix metaframe xpa fr3 servers
1 ms sql db-server (which holds the data store db)
1 web server with web interface 2.0 installed
What I want:
Load balancing & the ability to have some degree of fault tolerance.
The load balancing stuff looks pretty straight-forward (after trial and error). But there is still some mysteries about this web-interface product.
For instance, when I install the webinterface I had it point to my metaframeserver1 machine. Afterwards I went into the wiadmin interface and added metaframeserver2 as the second server to use (also as the second machine in the list order).
I do a test logon against both machines (not through the web-interface but through the ica-client) just to check that the servers hangs with me. Then I publish an application (notepad) and do a test logon against metaframeserver1 from the web interface (I guess it is the machine that answers, cause it is the first in the server-list).
Then, I power off the metaframe1 server and tries to log on once more. Nothing happens (should not server2 be the point of authentication right now?). I go to the webinterface admin console (wiadmin) and change the order of my servers so server2 is the first one and server1 is the second one. Try another log-in (same result).
Then I boot up metaframeserver1 again ... I even change back the server order so metaframeserver1 is first in the list, but it still fails to log on.
My impression is that the web interface seems a bit moody (be nice to me, or else I refuse to cooperate
) ). The strange part is that the failed logons do not produce any error-statement in the web-interface. It just pops back to the log-on screen with empty user, password and domain fields.I should be able to reach my farm even if one of my servers did not respond (for instance due to a power failure) right? And as long as I put the data store on a db-server (independent of the metaframe servers ... they should reach all the stuff they need without needing anything from each other?
Can anyone fill me in on this one?