[Swan-dev] setting up tests on a brand new Fedora 23 installation

[D. Hugh Redelmeier]

Much progress.  I ran one test successfully.

Since then, I updated libreswan and tried a full.  This did not work.

================

First updating libreswan (as per <https://libreswan.org/wiki/Test_Suite#Preparing_the_VMs>):

I built the software on west.  That worked.

I installed the built sofware on east north and road.  That worked
except for north.  It seemed to not come up in time so I rebooted it
and tried again and that worked.  Suspicious, but I ignored this.

================

I started the full test and went away for a few hours.  When I came
back, the tests had aborted.  I'm not sure what it was doing but the
root cause of the failure was that west did not come up fast enough
after a boot.

This was on the console:
[ *** ] A start job is running for Load Kernel Modules (4min 12s / no limit)(west [2] waiting)

Looking inside west, using "journaclctl -b" (meaning "show logs for
the current boot"), I see the following suspicious entries:

May 13 16:43:50 swanbase kernel: [drm] Initialized drm 1.1.0 20060810
May 13 16:43:51 swanbase kernel: [drm:qxl_pci_probe] *ERROR* qxl too old, doesn't support client_monitors_config, use xf86-video-qxl in user mode
May 13 16:43:51 swanbase kernel: qxl: probe of 0000:00:02.0 failed with error -22
May 13 16:44:51 swanbase systemd-udevd[257]: worker [264] /devices/pci0000:00/0000:00:07.0/virtio4 is taking a long time
May 13 16:45:32 swanbase kernel: random: nonblocking pool is initialized
May 13 16:46:51 swanbase systemd-udevd[257]: worker [264] /devices/pci0000:00/0000:00:07.0/virtio4 timeout; kill it
May 13 16:46:51 swanbase systemd-udevd[257]: seq 1399 '/devices/pci0000:00/0000:00:07.0/virtio4' killed
May 13 16:46:51 swanbase systemd-udevd[257]: worker [264] terminated by signal 9 (Killed)
May 13 17:14:17 swanbase systemd-modules-load[224]: Inserted module 'virtio_rng'

Notice the time delay between the last two messages.

This looks to be the problem.