Have had an issue where template based servers (rebuilt weekly) could go out of sync in ActiveDirectory, but shows normal/ok in RAS Console.
Last case instance today, where a host pool of 4 RDSH Terminalservers, all recreated from template sunday at 04:00, where a single host had trust relationship problems towards AD.
As Parallels Ras RDSH Agent was reporting all okay, the server was in full production and showing a big green OK, but would not log users on.
As this one server had a lot less users logged on, LoadBalancing would try to point every new session to this server.
Setting server in logon disabled state would let users log on, and recreating the host from template fixed the issue. Quick and simple.
My feature request is then to continously every "n hours" do authentication tests/logon tests and to set host in "failed" state/logon disable state if any reoccuring failures are heppening. Test could be done as simple as %server%\ipc$ or the whole RDSH protocol could be tested with a test login/logout.
Agent could parse the Eventlog for trust relationshop problems, authentication denied errors or similar.
Also, it would be great if the Parallels RAS RDSH Agent could monitor users login/session, and possibly redirect to other servers if login is failing to the first attempted server in the pool.
With multiple errors or with trust relationship problems detected, the single server could be automatically recreated from hostpool template.
As this is working per now, a failed server would render the whole host-pool in a "invisible" failure state if login errors occur du to the fact that no users are logged in to this server, and loadbalancing really want to even users out.
Last case instance today, where a host pool of 4 RDSH Terminalservers, all recreated from template sunday at 04:00, where a single host had trust relationship problems towards AD.
As Parallels Ras RDSH Agent was reporting all okay, the server was in full production and showing a big green OK, but would not log users on.
As this one server had a lot less users logged on, LoadBalancing would try to point every new session to this server.
Setting server in logon disabled state would let users log on, and recreating the host from template fixed the issue. Quick and simple.
My feature request is then to continously every "n hours" do authentication tests/logon tests and to set host in "failed" state/logon disable state if any reoccuring failures are heppening. Test could be done as simple as %server%\ipc$ or the whole RDSH protocol could be tested with a test login/logout.
Agent could parse the Eventlog for trust relationshop problems, authentication denied errors or similar.
Also, it would be great if the Parallels RAS RDSH Agent could monitor users login/session, and possibly redirect to other servers if login is failing to the first attempted server in the pool.
With multiple errors or with trust relationship problems detected, the single server could be automatically recreated from hostpool template.
As this is working per now, a failed server would render the whole host-pool in a "invisible" failure state if login errors occur du to the fact that no users are logged in to this server, and loadbalancing really want to even users out.
