Problem with notfred's bootcd

[tedkj]

Hi,

I've been messing around with the Dell 2850 servers, and set up VMWare Server 2 Beta 2 on W2K8 Server. All seems to work fine with VMWare, however, when I run the notFred's bootcd (or the premade drive-image), it atuo-runs the fold 32bit client.

If I override it and run fold64, it craps out at loading the kernal and says I need a x86-64 CPU and I only got an i1586 (?) CPU.

The W2K8 Server is 64bit, running on two DualCore 3.0 Xenons..

Should I just setup a 64bit version or Debian or Ubuntu and run the SMP's manually, or is there some easy way to fix the problem with notFred's system?

Any help highly appreciated. I'm gunning for the OCAU crowd!

 

[nomad8u]

A couple of things might be going on. First a question, are you running the GPU2 client while trying to start the VM's? Looking at your output/uploads on the stats page it appears so.

I was running GPU2 + a NF(notfred) VM SMP client and a Windows console client and wanted to experiment with 2 NF VM SMP clients with GPU2. I'm running a Q6600 @3.1 and a 9600GT @ 710/1825/1000.

I'm using the ISO download from notfred BTW. I went to the site looking for updates and had the June 9 build I downloaded that I was going to use for the 2nd VM SMP client. The 1st VM SMP client had and is using the March 26 build FWIW...

When I tried starting the 2nd client, I got the message about not having a 64bit processor, and since I was using the "fold64.vmx" it hung. So I stopped the GPU client and tried it again but had the same result, so I went back to using the 26 March build on the second client since I still had a copy configured for my setup.

The 2nd client still wouldn't start until I paused the GPU client (I'm guessing it couldn't find 2 free cores) but once I paused the GPU client, it started fine and I resumed the GPU client. It's been running almost 24hrs trouble free, I'm running VMWare Server v1.05 so I'm not using the same version you are.

My individual PPD for the 3 clients are down a little but overall PPD for the box are up about 600-700 over a GPU2+SMP+Console client mix. With the added benefit that the VM SMP's each eat 42-44% on the 4 cores, and keep the GPU2 client limited to 8-12% so I no longer have all the desktop/application lag from my GPU client... WU's are still WAY under preferred deadline so it's all good for now..

I may try a GPU client+VM SMP+VM standard client to see how that fares...

Since I ran on so much... (sorry ) I'd try 2 things for your problem.
1. Go setup and download another ISO from notfred in case the first one was corrupted somehow.
2. Try pausing your GPU2 client if it's running, until you get the VM started successfully, then resume the GPU2 client.

Good luck.... home some of the info helps.

 

[Tigerbiten]

I run the Linux client inside my VM's & the nVidia GPU2 client.
One thing I've found is that I get a B.S.O.D. every time if I start the GPU2 client before I start the VM's.
Start the all the VM's first and everything runs like a dream.

As for lag ..........
I've found that on my Dual-Quds running Windoze XP, if I isolate all the VM clients on 7 cores and run Windoze and the GPU2 client on the 8th core, I dont suffer from lag.
If I run it the other way around or dont set affinities then the lag is noticable.
I use SMP Seesaw Pro to isolate the first VM and set affinities of the others

Luck ......... .

 

[tedkj]

No, I got a rack of a row of Dell 2850 19" servers which I'm trying to run the VM on.. The GPU2 clients I got running currently chew on a GTX280 and GT8800 on my normal boxes.