B@D Known Issues, Workarounds and Troubleshooting Guide

This page details the known issues and any workarounds when working with the B@D VM. The issues are tracked in the CIP bugtracker on gitlab.com. Links to the issues are given when possible. Additionally, the CIP testing team has started compiling a troubleshooting guide for some common problems we've seen that are not bugs.

Known issues of the VM created within the CIP project including LAVAv2 and KernelCI.

LAVA looks for the U-Boot prompt but due to differing forms of this prompt may not find it, see https://gitlab.com/cip-project/cip-testing/testing/issues/53 for background and a manual fix for the beaglebone.

If you suspend your host with Vagrant still running, when you awaken it the VM still has the time when suspended and it does not get updated to the time now. See https://gitlab.com/cip-project/cip-testing/testing/issues/87 for the associated issue. With Vagrant 2.1.2 this issue no longer occurs so upgrading your vagrant installation is a workaround - and therefore that issue has been closed.

If the QEMU Health check times out, try adjusting the values in the script along the lines of those values mentioned in https://gitlab.com/cip-project/cip-testing/testing/issues/121

If during “vagrant up” the VM becomes unresponsive at the “Configuring and enabling network interfaces” step.

Press Ctrl-C twice to force a shutdown of the VM and restart it by issuing the vagrant up command again. Normal vagrant control should be possible after the restart.

Note: You can also force a shutdown using the Virtualbox GUI.

Below are a list of errors you might see during the vagrant up process:

A. GetPassWarning: This warning is safe to ignore on our Board-at-Desk Single-Developer VM. It does not affect the operation of the KernelCI/LAVA VM.

1. GetPassWarning: Can not control echo on the terminal." - "Warning: Password input may be echoed."

B. CommaSeparatedIntegerField: This warning is safe to ignore on our Board-at-Desk Single-Developer VM. It does not affect the operation of the KernelCI/LAVA VM.

1. ==> default: lava_scheduler_app.Notification.job_status_trigger: (fields.W901) CommaSeparatedIntegerField has been deprecated. Support for it (except in historical migrations) will be removed in Django 2.0.
2. ==> default: HINT: Use CharField(validators=[validate_comma_separated_integer_list]) instead.

The Vagrant instance by default is configured with 2GB of memory, if you require more memory, and have sufficient local resources on your machine, edit line 26 of the Vagrantfile so that vbox.customize [“modifyvm”, :id, “–memory”, “2048”] is given a larger value.

Similarly, if you need a larger disk storage capacity, there are several ways you can accomplish that. The preferred method is shown below.

1. To Add a new hard drive to the B@D Virtual Machine, shutdown the VM using vagrant halt http://www.techrepublic.com/article/how-to-add-new-drives-to-a-virtualbox-virtual-machine/

2. Using the GUI, add a hard drive to the VM (/dev/sda)

3. Boot the VM using vagrant up

4. Create a new partition table on the new disk using fdisk https://askubuntu.com/questions/384062/how-do-i-create-and-tune-an-ext4-partition-from-the-command-line

5. Create a new primary partition (/dev/sda1)

6. Write your changes to disk

7. Make a directory where you want to mount the new disk to (/storage)

8. Mount /dev/sdb1 to /storage

9. Consider moving the /var directory to the new drive since that is where the build artifacts are stored. https://linuxconfig.org/how-to-move-var-directory-to-another-partition. As the /home directory is likely to be larger than /var unless you do a lot of builds of different versions of the CIP kernel you may also wish to move /home onto its own partition.

(If wanting to expand primary vmdk disk sda, resizing via vdi & gparted: https://computingforgeeks.com/step-by-step-guide-on-how-to-extend-vdi-and-vmdk-hard-disks-on-virtualbox/)

If you have not rebooted the virtual machine after the initial provision, logging in as administrator will return you to the home page but with the login still available. You need:

vagrant halt 
vagrant up

in order to have the LAVA installation running cleanly so that login works.

In order to have versions of these packages which work together in Debian stretch, run the following (on the host!) see #121:

sudo apt-add-repository 'deb http://download.virtualbox.org/virtualbox/debian stretch contrib'
curl -O https://www.virtualbox.org/download/oracle_vbox_2016.asc
sudo apt-key add oracle_vbox_2016.asc
sudo apt-get update
sudo apt-get install virtualbox-5.1

If you have problems connecting to the BBB using ssh - though the health check does not reuire this - use the following procedure:

• The usb power cable must be connected to the host (as well as the ftdi cable but this isn't used here) the eth1 device will appear - but is not being assigned an IP address
• Then sudo ifconfig eth1 192.168.7.1 (on the host) ssh root@192.168.7.2 should then work.
• Note that the BBB will reboot during the health test and the above setting will be lost so an edit to /etc/network/interfaces will be necessary to set this address

Builds in kernelci become invisible to the web interface after 14 days - though the builds are still accessible from LAVA tests, if you wish to increase that timeout edit the DATE_RANGE setting in /etc/linaro/kernelci-frontend.cfg. See the discussion in #164 - now closed.

You may have errors when trying to provision B@D when operating behind some web proxies, see #195 for a discussion and hints.

Back to the B@D Features page