ubuntu_performance_multipass: Repeated failure to launch VM

[Francis Ginther]

Typically we see the multipass launch command fail on the first few attempts. Typically, this is ignored (don't know why), but could lead to errors later when calculating stddev: https://bugs.launchpad.net/ubuntu-kernel-tests/+bug/2067978

Here's a log from a PASSING run:

12:22:47 INFO | START ubuntu_performance_multipass.multipassboot ubuntu_performance_multipass.multipassboot timestamp=1717762967 timeout=1800 localtime=Jun 07 12:22:47
12:22:47 DEBUG| Persistent state client._record_indent now set to 2
12:22:47 DEBUG| Persistent state client.unexpected_reboot now set to ('ubuntu_performance_multipass.multipassboot', 'ubuntu_performance_multipass.multipassboot')
12:22:47 DEBUG| Waiting for pid 7287 for 1800 seconds
12:22:47 WARNI| System python is too old, crash handling disabled
12:22:48 INFO | failed to launch
12:22:48 INFO | failed to launch
12:24:25 DEBUG| Running 'multipass list'
12:24:25 DEBUG| [stdout] Name State IPv4 Image
12:24:25 DEBUG| [stdout] test2xenial Running 10.229.31.50 Ubuntu 16.04 LTS
12:24:25 DEBUG| Running 'multipass exec test2xenial -- uname -r'
12:24:26 DEBUG| [stdout] 4.4.0-210-generic
12:24:26 DEBUG| Running 'multipass exec test2xenial -- systemd-analyze'
12:24:26 DEBUG| [stdout] Startup finished in 3.362s (kernel) + 30.160s (userspace) = 33.522s
12:25:38 DEBUG| Running 'multipass list'
12:25:38 DEBUG| [stdout] Name State IPv4 Image
12:25:38 DEBUG| [stdout] test3xenial Running 10.229.31.61 Ubuntu 16.04 LTS
12:25:38 DEBUG| Running 'multipass exec test3xenial -- uname -r'
12:25:38 DEBUG| [stdout] 4.4.0-210-generic
12:25:38 DEBUG| Running 'multipass exec test3xenial -- systemd-analyze'
12:25:39 DEBUG| [stdout] Startup finished in 3.345s (kernel) + 39.362s (userspace) = 42.708s
12:26:43 DEBUG| Running 'multipass list'
12:26:43 DEBUG| [stdout] Name State IPv4 Image
12:26:43 DEBUG| [stdout] test4xenial Running 10.229.31.129 Ubuntu 16.04 LTS
12:26:43 DEBUG| Running 'multipass exec test4xenial -- uname -r'
12:26:43 DEBUG| [stdout] 4.4.0-210-generic
12:26:43 DEBUG| Running 'multipass exec test4xenial -- systemd-analyze'
12:26:44 DEBUG| [stdout] Startup finished in 3.521s (kernel) + 30.785s (userspace) = 34.306s

Notice the two "failed to launch" messages. This is coming directly from the code to launch the VM:
    def multipass_boot(self, vm, release):
        cmd = 'multipass launch -v -c 8 -m 1G -n ' + vm + ' daily:' + release
        if self.multipass_run_cmd(cmd):
            print("failed to launch")
            return None

Why does this happen so often? Is it masking a problem? Do we need to wait longer or prep the machine more or keep retrying these?