E
E
Eugene2015-10-19 16:32:54
VPN
Eugene, 2015-10-19 16:32:54

How to improve the stability of RDP connection?

There are two locations. Between them, an ipsec tunnel has been raised between public addresses using public channels (Internet). Different traffic goes in the tunnel, mainly VoIP and RDP.
Problem - RDP connections to terminal servers (a cluster with a broker) often (several times a day for a user) are reconnected. The tunnel itself remains intact. Event ID 4779 and 4778 appear in the terminal server logs, the difference between events is on average 5-6 seconds.
What settings can make an RDP session "more tenacious"? Disable compression (channel allows) tried, did not help.
PS: the problem was different, I take out the solution from the comment, which is not very noticeable:
As a result, the problem turned out to be in a completely different place. A long time ago, my predecessors set up a 10 minute address issuance time on the dhcp server (many wifi clients, small pool of addresses). It worked for years and there were no special problems. But with my arrival, they began to change the architecture, added a dhcp relay, and problems with RDP reconnects started. Logs on terminal servers showed that particularly unfortunate users were reconnecting every 10 minutes, clearly (but not all).
It turned out that the addition of relay increased the travel time for packets and cases began to appear when the host had already expired and had not yet received a new one (more precisely, the old address, but again). It was in these fractions of a second that reconnects occurred, somewhere more often, somewhere rarely.
Another factor that initially hindered diagnostics - the problem manifested itself only on stations with Win7, machines with Win XP on adjacent ports of the switches did not experience any problems at all.
Now Lisa has set the time to 8 days, the problem is fixed.

Answer the question

In order to leave comments, you need to log in

2 answer(s)
E
Eugene, 2015-12-17
@misant

As a result, the problem turned out to be in a completely different place. A long time ago, my predecessors set up a 10 minute address issuance time on the dhcp server (many wifi clients, small pool of addresses). It worked for years and there were no special problems. But with my arrival, they began to change the architecture, added a dhcp relay, and problems with RDP reconnects started. Logs on terminal servers showed that particularly unfortunate users were reconnecting every 10 minutes, clearly (but not all).
It turned out that the addition of relay increased the travel time for packets and cases began to appear when the host had already expired and had not yet received a new one (more precisely, the old address, but again). It was in these fractions of a second that reconnects occurred, somewhere more often, somewhere rarely.
Another factor that initially hindered diagnostics - the problem manifested itself only on stations with Win7, machines with Win XP on adjacent ports of the switches did not experience any problems at all.
Now Lisa has set the time to 8 days, the problem is fixed.

S
Saiputdin Omarov, 2015-10-19
@generalx

Ha, we use the wan optimizer - riverbed

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question