Terminal server immediately disconnects

IndieSchmindie · 15 Jan 2010 at 14:47

Hi all,

I've been looking round the net for similar problems without a lot of success.

We have a server at work that periodically starts dropping all RDP connections. There is no error message, the screen starts to open then immediately closes before even prompting for a username/password.

It happens from all PC's and I've tested using a few different TS clients.

The only solution is to reboot the server, but that is a real pain because it is in a production environment.

The ports are open and listening - I guess if they weren't it would return an error anyway.

Any suggestions gratefully received.

Yamahahahahaha · 15 Jan 2010 at 17:21

What does event viewer say?

GhostlyPea · 15 Jan 2010 at 17:29

Had a similar problem on Citrix, turned out o be a liscencing issue

- Pea0n

Uhtred · 15 Jan 2010 at 19:28

We had the same problem on 2003 with an HP server. Seemed to be related to VSS but we never really got to the bottom of it. After we turned shadow copies down from 1 hour intervals to 4 hours the problem happened a lot less often.

Not sure if that applies to you but thought i'd mention it.

iaind · 15 Jan 2010 at 21:02

If the problem reduced when you lengthened the time between VSS snapshots, it's probably a duff RAID controller/driver etc

This could be a licensing issue though

MagicBoy · 15 Jan 2010 at 21:07

Licensing methinks.

iaind · 16 Jan 2010 at 11:50

Well the general concensus is a possible licensing issue, lets expand on that.

Where is the licensing - on the TS, a DC or elsewhere? User or Device CALs? Are you using discovery or manually specifying a license server? Is the license service running - can you view the licenses in TS license manager when the issue occurs? Try clearing the license cache from the client and see if the problem occurs again.

I've also seen similar things with "black hole" routers, but that generally manifests itself in a black screen rather than a disconnection - if they're on a WAN it may be worth playing with MTU sizes but licensing would be my first port of call

GhostlyPea · 16 Jan 2010 at 15:03

To expand on the issue we had, the licensing server was not on the citrix server. The licensing server was the problem and we needed to update it, once done all was fine so the problem wasn't with citrix itself.

- Pea0n

IndieSchmindie · 16 Jan 2010 at 19:08

Sorry I should have said that it's not a Terminal server in the sense that its in Application mode. It's only use for administration and therefore isn't a licensing problem.

Interestingly, Shadow Volume copies is on, but only twice a day. However there is no real need for it so I've disabled and will see if it helps. The problem occurs normally about once a week on average so we'll see.

Event view has nothing by the way.

IndieSchmindie · 16 Jan 2010 at 19:10

Should have mentioned the hardware as well! It's an HS21 blade in a BladeCentre S

Trojan · 17 Jan 2010 at 00:50

Are you able to log onto the server locally when the disconnects occur?

IndieSchmindie · 17 Jan 2010 at 01:11

.Simon · 19 Jan 2010 at 11:24

Could it be down to some policy settings in gpedit or secpol? Timeout settings?

IndieSchmindie · 19 Jan 2010 at 14:29

Don't think so. I've got nothing configured in GP that would cause that I don't think you can set the timeout to 0.25 seconds

Uhtred · 21 Jan 2010 at 20:29

Has it dropped at all this week?

IndieSchmindie · 24 Jan 2010 at 19:17

Not yet! Looking like the shadow volume copy. Is there a hotfix available?

Uhtred · 26 Jan 2010 at 11:16

IndieSchmindie said:
Not yet! Looking like the shadow volume copy. Is there a hotfix available?

Not that I know of. For us it just seemed to go away but I would do as iaind suggested and check your RAID controller drivers etc.

IndieSchmindie · 28 Jan 2010 at 00:47

It still hasn't done it so I think this is the issue.

It's a fairly recent build done with the latest IBM ISO downloaded from their site rather than that which came with the server.

sparkey196 · 2 Feb 2010 at 11:26

I had a similar problem which affected only certain clients. It turned out to be caused by a packet fragmentation issue, the details of which I cannot remember, tweaking the vpn settings fixed the problem. Something to consider.

IndieSchmindie · 3 Feb 2010 at 22:36

Thanks.

It hasn't done it since I turned off shodow copies it's been fine so I think that was the problem. I'm guessing there is some kind of bug with it, or with the raid controllers for IBM HS21 blades.