Would adding --debug require a change to tools/lxd/client.go? If so, can you provide an example?
I am unsure whether the machine in Cheryl's example is overloaded. The machines in my example were starting 8 lxd instances.
Regarding decoding the logs,
Basically, two attempts are run simultaneously, they are named "parallel-azure-old" and "parallel-azure-new" for historical reasons and they are given odd and even numbers respectively.
Would adding --debug require a change to tools/lxd/ client. go? If so, can you provide an example?
I am unsure whether the machine in Cheryl's example is overloaded. The machines in my example were starting 8 lxd instances.
Regarding decoding the logs,
Basically, two attempts are run simultaneously, they are named "parallel- azure-old" and "parallel- azure-new" for historical reasons and they are given odd and even numbers respectively.
Combining the fact that 0/lxd/2 failed for the first attempt, we can say that the files linked here describe a forkstart failure: juju-ci. vapour. ws/job/ industrial- test-azure/ 126/artifact/ logs/0/ parallel- azure-old/ machine- 0/
http://