Fetching latest commit…
Cannot retrieve the latest commit at this time.
|Failed to load latest commit information.|
Integration test for CloudBioLinux. When you have a base install running, you only need to supply the (local) IP addrees of the VM, and the password-less key for the install user (who has sudo bash). The tests get invoked wiht ./test_biolinux [-i key.pub] [-u user] IPaddr [Flavor] For further options see ./test_biolinux --help Alternatively try the self running test using Virtualbox and Vagrant with: ./test_vagrant --help == Testing issues with Vagrant == The test system depends on a stack of tools, including virtualbox, vagrant, ssh, ruby and related libraries. Things can go wrong. Here we list a number of common problems. === Error with ssh config === The test system appends login code to enter a running vagrant virtualbox without asking for a password. Sometimes the contents of $HOME/.ssh/config get corrupted. Check the contents out, and remove all wrong information with vi ~/.ssh/config and rerun the test. === Error suggests 'vagrant up' === If something went wrong with the install, Vagrant may complain that you need to run 'vagrant up' first. In practise, this may mean: * Remove other running vagrant instances (vagrant halt) * Remove an existing (faulty) instance using virtualbox * Check ~/.ssh/config contains no faulty lines * Rerun the tests from scratch, i.e. without the '--continue' switch === Fatal error: Name lookup failed for vagrant Probably there is another VirtualBox + vagrant running somewhere. This conflicts with the information in ./ssh/config. Easy solution: remove and kill the running instance first. === Error: ffi.rb:106:in `call_and_check': Error in API call to get_state === This is Ruby trying into the vagrant API. You need to reinstall the vagrant and virtualbox gems to match your version of Ruby. === Grub error On a kernel upgrade, the test scripts may want to install grub, which is the boot loader. Vagrant does not need that, so don't try to install (select ignore).