ILO 3 no Java remote console keyboard response in RHEL5

[W.C. Epperson]

Due to network constraints, I need to be able to do ILO remote console via a RHEL5 box. I've installed 32bit Firefox 3.6.17 and 32bit Java jre1.6.0_25 and done a shell script to run them:

#!/bin/bash
export FF32=/opt/firefox32
export JAVA_HOME=$FF32/jre
export MOZ_PLUGIN_PATH=$FF32/firefox/plugins
export PATH=$FF32/firefox:$FF32/jre/bin:$PATH
$FF32/firefox/firefox

I get to the ILO gui just fine. When I launch the Java Integrated Remote Console, I get the screen painted and can see the login prompt, but the keyboard doesn't do anything when I type. The mouse is active in the console window, and I can use the power, virtual drive, and keyboard tabs at top. I found a post about disabling persistent mouse and keyboard under Power Management-->Power Settings and did that. The only change was that now when I launch the Java IRC, I see a new "usb usb6: Controller not stopped yet!" message. I.E. after 1 launch I saw one such message on the console, after 2 I saw 2, etc.

DL380 G7 with ILO 3 v1.16 firmware

I have a bunch of these, so I really don't want to do it with port forwards and a Windows box.

[Oscar A. Perez]

Does it happen every time you open the Java RC? I've seen cases where the keyboard/mouse doesn't the first time I open RC but closing the RC session and opening it again seemed to solve it.

Also, by disabling the persistent KB/Mouse you are going to cause PnP events every time you open close RC. It is almost like having an external USB KB and Mouse that you manually attach to the server every time you open RC and de-attach every time you close RC. If Linux USB driver has a problem and it's not handling the USB PnP events correctly then, the problem could be in your OS. Try rebooting the OS and see if you get your KB/Mouse back.

[W.C. Epperson]

Yes, it does happen every time I open the Java RC.

I disabled the persistent KB/Mouse based on an HP technical article on the issue. The only impact I could see was the "Controller not stopped" messages.

Rebooting the OS did not get my never-working KB "back". It did stop the "Controller not stopped" messages from printing on console every time I open the Java RC.

Thanks for your suggestions and questions. I'm willing to try anything else, within reason. This particular machine is not yet handed off to customer, so I can reboot at will.

[Oscar A. Perez]

When you reboot the system with RC open, does the KB work during POST? I mean, if you press F9, will it go into System ROM Conf Utility?

[W.C. Epperson]

I'll check. Right now I have the persistent kb/mouse in "Enabled" state. Should I conduct further tests with that enabled, or disabled? These things take foreeeeever to post up, so I don't want to have to do two reboots per test.

[W.C. Epperson]

General question: has anybody _ever_ gotten an ILO 3 Java RC to work properly from Firefox 3.6 on a RHEL5 x86_64 platform by installing 32 bit FF/Java? If this horse is dead, no point in beating it.

[Oscar A. Perez]

Keep persistent KB/Mouse enabled. If the KB works during POST then, the problem is in your OS (maybe not seeing/configuring the iLO3 UHCI controller).
If the KB doesn't work during POST then, you probably disabled some USB setting in the System ROM and you are going to reverse those settings back to default.

[W.C. Epperson]

Nope, it ignored F9 and went on into OS boot.

[Oscar A. Perez]

Did you disable some of the USB settings in the ROM?

[W.C. Epperson]

You mean in the BIOS setup? No, it ignored the F9, so it didn't go into setup.

What particular settings are you thinking of? I'll have to either go to the data center (30 miles) or rig up a port forwarding scheme to get to the BIOS settings. Unless there's a

Oh, I realized that with persistent kb/mouse=enabled, I just get BIOS/post video, nothing at all once OS takes over. With disabled, I see the OS console output.