VMware Horizon Community
mittim12
Immortal
Immortal

Connection Closed

I had two users today who received the error message "The connection to the remote computer ended" when trying to login to a VDM 2.1 non persistent pool. This was a small pool so I was able scan the agent logs until I found the problem computers.. My problem with this is the connection broker did not provide any usefull information for trying to determine which desktops were the problems. The only error message I received from the Connection Broker was the one below.

FE1262D38ACF734B9441109F3266B47A) Conveyor 0 IOException: java.io.IOException: An existing connection was forcibly closed by the remote host com.vmware.vdi.ice.server.j.a(SourceFile:337)

java.io.IOException: An existing connection was forcibly closed by the remote host

I am also concerned with the connection broker attempted to route users to desktops that were obviously generating error messages from the VDM agent. Has anyone else experienced this? Agent logs are below.

06:55:41,547 WARN <MessageFrameWorkShare> [] Java HotSpot(TM) Client VM warning:

06:55:41,547 WARN <MessageFrameWorkShare> [] Attempt to allocate stack guard pages failed.

0 Kudos
2 Replies
mpryor
Commander
Commander

(FE1262D38ACF734B9441109F3266B47A) Conveyor 0 IOException: java.io.IOException: An existing connection was forcibly closed by the remote host com.vmware.vdi.ice.server.j.a(SourceFile:337)*

This error will occur when the guest VM terminates the tunnelled TCP connection (most likely the RDP connection) with the connection server. I can agree with you that it's hard to match up to the specific desktop unless you know what you're looking for. The 32 character string is the tunnel session, it should be possible to track this back up through the logs to see who it belongs to and what it has done.

WARN <MessageFrameWorkShare> \[\] Java HotSpot(TM) Client VM warning:

WARN <MessageFrameWorkShare> \[\] Attempt to allocate stack guard pages failed.

Hard to guess what is causing this or whether it's a problem that was present while the user was connecting. Could you provide a support bundle on this VM? support.bat in the DCT subdirectory will generate it.

0 Kudos
mittim12
Immortal
Immortal

I will try to get the log files but my biggest complaint is the fact that it was still displaying a "ready" state to the broker so the broker continued to attempt to route people to it. Not sure how the broker chooses which desktops to route people to but it also seemed to be the next line so it was if these users continued to hit this one even though there were other available machines in the pool.

0 Kudos