Nested RDP and VMWare and ILO console sessions: keystroke repetition and latency
Whereas an SSH connection transmits keystrokes, an HP ILO connection transmits key states. Each time you press a key, the server receives separate KeyDown and KeyUp events. The repeated keystrokes result when the KeyUp event is received late.
The two most likely reasons for the KeyUp event to be received late are:
- Network congestion/performance issues.
- Poor performance of the client system initiating the ILO connection. If the client is a virtual machine, is the underlying host system overloaded, or does the VM have inadequate memory/CPU resources allocated?
If the root cause cannot be addressed:
- The key repeat problem can be worked around by disabling an ILO2 setting called "Key Up/Down". This will cause ILO2 to transmit keystrokes instead of key states. Unfortunately, this setting was removed from ILO3.
- If the target operating system is Linux, you may be able to work around the issue by redirecting the console to
ttyS0and using a Virtual Serial Port (VSP) session instead of a virtual console. This will eliminate the Key Up/Down issue, because serial connections transmit keystrokes instead of key up/down events.
- It may be helpful to adjust the key repeat rate and/or disable autorepeat entirely on the target system. I acknowledge that this may not be easy to accomplish, depending on the severity of the key repeat problem.
- Given that you're using a Mac as your local workstation, it might be worthwhile to try pasting complete commands into your Mac RDP client using Command-V. I do not know whether this is a viable workaround, but it might have an interesting effect. I've often appreciated working on remote Windows machines from a Mac workstation specifically because local command-hotkey combinations continue to work predictably.
- Advisory: HP Integrated Lights-Out 3 (iLO 3) - Using iLO 3 Integrated Remote Console (IRC) Over a Slow or Congested Network May Cause Keystrokes to Repeat
- HP Communities - ILO3 typing / mouse lag problems?
This looks likes it's just a problem with the protocol. I've reduced the issue somewhat by using Ericom Blaze as the RDP transport for the central server I connect from; e.g. "jump box".
I'm trying to avoid multiple nested sessions.
I'm running VMWare Fusion with Windows 7 on my Mac to allow me to use the native RDP from Windows in certain cases.
That's about all I can see for now.
you need to edit the .vmx file to add the following line:
keyboard.typematicMinDelay = "2000000"
it takes out the "bounce".
With my version of vmware, I have to make this change when the VM is down. I understand that it can be made from an edit window, but I have not been able to find that place.