Folding@Home Header

How to deploy the VMware Appliance for Folding@Home

In this blog post we will go through the steps to deploy the VMware Appliance for Folding@Home to;

  • vCenter
  • Standalone ESXi host
  • VMware Fusion/Workstation

And also cover some basic troubleshooting.

Getting started with the VMware Folding@Home Appliance (vBrownBag Recording)

Deploy the VMware Appliance for Folding@Home to vCenter

  1. Select “Deploy OVF Template” in vCenter
  2. Input your virtual machine name, and select the folder, compute resource.
  3. Review and accept the template details
  4. Select your datastore
  5. Select your virtual machine destination network
  6. Configure the appliance properties
  • Hostname
  • IP Address
    • Note: If no IP address is set, appliance will use DHCP
  • Network CIDR Prefix
  • Gateway
  • DNS
  • DNS Domain
  • NTP
  • Proxy Settings
  • OS Credentials – Root password
    • Note: this is a mandatory parameter, failure to set a password will lead to a failed deployment.
  • F@H Mode
  • GPU enabled
    • Note: if you are using a virtual machine with a GPU, this must be in passthrough mode
  • F@H Remote Mgmt allowed networks
    • Note: Default allows all connections from FAHControl
    • 127.0.0.1 must always be present, accept inputs are [IP Address] [IP Network/CIDR}
    • Example;
      • 127.0.0.1 192.168.10.10
      • 127.0.0.1 192.168.10.0/24
  • F@H Remote Mgmt password
    • Note: This sets a password to be used when connecting FAHControl to your Appliance. The default is VMware1!
  • Enable F@H Stats in VM Console

For appliance version 1.0.1 and higher, there has been a change which allows the auto configuration of “Big Advanced” workloads when your appliance detect 16 CPUs or higher. This detection will run upon every boot. You can read more about this setting here;

Deploying the OVA directly to a ESXi Host

Deploying the OVA to an ESXi host using the host client UI is not supported, you must instead use the OVFTool;

  1. Create the script for your system, you will find options for Windows/Linux/MAC OS X here;
  2. Modify the script with variables for your system in a code editor
  3. Modify the script to be executable
    • For Mac OS X and Linux;
      • Open terminal “chmod +x {script location}
      • chmod script to deploy vmware appliance for foldingathome
  4. Run the script

script output to deploy vmware appliance for foldingathome

Once the script has finished, you will see a deployed and powered on virtual machine that you can connect to using the web interface (https://{OVA_IP}:7396) or FAHControl.

Deploying OVA to VMware Fusion & VMware Workstation

This is supported for OVA version 1.0.1 and higher.

Troubleshooting

I’m not receiving any work units?

If you think there is something wrong with the appliance,

  1. Check the status of the Service;
/etc/init.d/FAHClient status
  1. You can restart the services with;
/etc/init.d/FAHClient stop

/etc/init.d/FAHClient start

Or

/etc/init.d/FAHClient restart

You can then view the logs as below

Logs Location on OVA:

Run /etc/init.d/FAHClient log -v

Or

Run less /var/lib/fahclient/log.txt

If you see logging as below, this means your appliance is actively running a work unit;

FAHClient logs showing running work unit

You can find the full Linux command line options at this website;

https://foldingathome.org/support/faq/installation-guides/linux/command-line-options/

Using FAHControl to manage multiple clients does not work

FAHControl uses the default TCP Port 36330

Test access with telnet you should get a response as below.

FAHClient telnet port 36330

The VMware Appliance for Folding@Home has IPTables configured to allow this port by default, if you did not specify a specific remote management address during deployment, then access is open to all IP addresses.

Ensure that the machine where you are running FAHControl is not blocking outbound connections to TCP 33630.

Regards

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.