If MAAS server lacks LXD images, deployments fail when using our preseed

[Rod Smith]

I tried deploying to a MAAS server that did not have LXD images installed locally, and the deployment failed, with the following complaint:

Setting up to use KVM image on MAAS server...
--2017-09-28 20:26:35-- http://172.24.124.1/cloud/
Connecting to 172.24.124.1:80... connected.
HTTP request sent, awaiting response... 404 Not Found
2017-09-28 20:26:35 ERROR 404: Not Found.

grep: /tmp/index.html: No such file or directory
grep: /tmp/index.html: No such file or directory
Setting up to use LXD images on MAAS server...
--2017-09-28 20:26:36-- http://172.24.124.1/cloud/
Connecting to 172.24.124.1:80... connected.
HTTP request sent, awaiting response... 404 Not Found
2017-09-28 20:26:36 ERROR 404: Not Found.

grep: /tmp/index.html: No such file or directory
grep: /tmp/index.html: No such file or directory
Setting up iperf...
Done.
Setting ubuntu password...
Done.
sudo: unable to resolve host brennan: Connection timed out
sudo: unable to resolve host brennan: Connection timed out
umount: /tmp/tmpgvdyakna/target/dev: target is busy
        (In some cases useful info about processes that
         use the device is found by lsof(8) or fuser(1).)
Unexpected error while running command.
Command: ['umount', '/tmp/tmpgvdyakna/target/dev']
Exit code: 32
Reason: -
Stdout: ''
Stderr: ''
curtin: Installation failed with exception: Unexpected error while running command.
Command: ['curtin', 'in-target', '--', 'sh', '-c', '# This script is looking for the SecureID in the following syntax as a tag in MAAS: secure_id-a00d000000abcde\n#!/bin/sh\nIFS=,\nfor i in `echo \'[\'secure_id-a00D000000dpNfPIAU\']\'`\n do\n if echo $i |grep -q "secure_id-"\n then secureid=$(echo $i|awk -F \'secure_id-\' \'{print $2}\'|sed -e \'s/\\[//g;s/\\]//g\')\n sudo sed -i "s/#\\[transport:c3\\]/\\[transport:c3\\]/g" /etc/xdg/canonical-certification.conf\n sudo sed -i "s/#secure_id =.*/secure_id = $secureid/g" /etc/xdg/canonical-certification.conf\n fi\ndone;\nsleep 5;\n']
Exit code: 3
Reason: -
Stdout: sudo: unable to resolve host brennan: Connection timed out
        sudo: unable to resolve host brennan: Connection timed out
        umount: /tmp/tmpgvdyakna/target/dev: target is busy
                (In some cases useful info about processes that
                 use the device is found by lsof(8) or fuser(1).)
        Unexpected error while running command.

I tried running "sudo maniacs-setup -d", but it didn't help, so maybe it's something else going wrong. I'm filing this bug report as a reminder to investigate this further and because it looks like a bug in our preseeds in either event.

[Rod Smith (rodsmith)]

The deployment failure seems to have been caused by a sporadic failure on my MAAS server -- maybe a corrupted file that was auto-updated last night. I can deploy fine now, even with the virtualization image directory renamed on the MAAS server. (The result is that /etc/xdg/canonical-certification.conf leaves the LXD and KVM image pointers commented out.)