For about 3 weeks now we have been getting random sessions disconnections
across our seven terminal servers (all windows 2003 sp2) They are all IBM
Blades. Remote users are only affected - it is not users within the local
lan that seem to be affected. Remote users connect in via a Managed MPLS VPN
Network.
On the MPLS VPN network we are losing packets. We are only losing packets at
the sites that connect into the MPLS via ADSL. We have one site that
connects into this via SDSL which no user has reported problems from. Our
head office connects to the MPLS via a 2mpbs leased line. Our network
provider has said that everything is within limits, and they say the don't
have any problems. All servers are on the lan within a gigabit backbone (at
head office). All remote sites use thin clients. There are no servers at
remote sites all at Head Office.
Users are disconnected when they actually are using their terminal. The
session will freeze and then terminal will re-connect back to the logon
screen. No user has reported any error messages and they have not reported
problems when connecting to a new session. Although we have had a few
corrupted profiles lately (which were reported in the eventlog). We don't
use roaming profiles (I have wanted to set these up but time has never
allowed). The corrupted profiles started after a message on the server
saying "Windows - Low on Registry Space the system has reached the maximum
size allowed for the system part of the registry. additional storage
requests will be ignored". We do have UPHC clean running on all TS as well.
I have tried setting the group policy setting for keepAlive to see if that
helps. On changing this one user reported a slight improve another
disconnected almost straightaway!
I have already posted this on the terminal services group. They believe this
is a network issue, and suggested I post here.
Any suggestion would be very helpful
Thanks
|