From a newsgroup post this error: "Citrix tends to start dying around a 700 ms latency (RDP is around 600ms), even at this latency, it starts to become unusable. Although it is not ideal, measure the latency using ping from his client machine to the Citrix server. If the latency is around 1000 ms, then almost certianly the session will die if data is streaming to the client at the time (high latency is OK, just so long as no data is being send to and from the server at the time of the hight latency).
Future release 1 for MF 1.8 has a "high latency" workaround which is essentially a local echo on the client of mouse and keystrokes - which makes most applications usable even at 700 ms latency - but it does not prevent the session from dying completely when the latency is 1000 ms or more. You will find that cable users will generally have 1/2 to 2/3 the latency compared to an equivalent DSL user. Note that to make the session stable, you only need to get the latency below 650 ms, rather than reduce it to a really low value.
Generally, there is not much you can do to lower the latency for the user - in these instances you need to look at the viability of shifting them to another medium such as cable, ISDN or even modems (we have had an office of 3 users running Citrix sessions concurrently very successfully on a 33.6k connection in the past - yes, we were suprised it worked at all)."
Another newsgroup post: "Solution: If the group Everyone is given Read access to the Mfc42u.dll file, the errors no longer appear during logon."
See Citrix document CTX891099
- This event is usually associated with a connection that gets disconnected early in the logon process. It usually occurs during the time that cltmgr queries the local client to determine what client devices are attached to the local client.