Print bootkube journal if it fails

[invidian]

In previous attempt in e852117, I
missed the fact that Terraform remote-exec only prints stderr to the
user and stdout is ignored. This means, if we redirect journalctl output
to stderr, then it should be printed which will improve debugging
experience.

Refs #440

Signed-off-by: Mateusz Gozdek mateusz@kinvolk.io

[invidian]

Cool, it seems this is working 🎉

module.aws-ci1604702277-hz.null_resource.bootkube-start: Creating...
module.aws-ci1604702277-hz.null_resource.bootkube-start: Provisioning with 'file'...
module.aws-ci1604702277-hz.null_resource.bootkube-start: Provisioning with 'remote-exec'...
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): Connecting to remote host via SSH...
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec):   Host: 18.184.225.254
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec):   User: core
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec):   Password: false
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec):   Private key: false
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec):   Certificate: false
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec):   SSH Agent: true
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec):   Checking Host Key: false
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): Connected!
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): Job for bootkube.service failed because the control process exited with error code.
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): See "systemctl status bootkube.service" and "journalctl -xe" for details.
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): -- Logs begin at Fri 2020-11-06 22:39:49 UTC, end at Fri 2020-11-06 22:42:24 UTC. --
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): Nov 06 22:42:22 ip-10-0-1-23 systemd[1]: Starting Bootstrap a Kubernetes cluster...
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): Nov 06 22:42:23 ip-10-0-1-23 bootkube-start[3304]: v1.19.3: Pulling from poseidon/kubelet
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): Nov 06 22:42:23 ip-10-0-1-23 bootkube-start[3304]: Digest: sha256:159718ffe057b1ccc092a1ae841c7bb96dd48daf9400d45219d7b484a71604fb
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): Nov 06 22:42:23 ip-10-0-1-23 bootkube-start[3304]: Status: Image is up to date for quay.io/poseidon/kubelet:v1.19.3
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): Nov 06 22:42:23 ip-10-0-1-23 bootkube-start[3304]: quay.io/poseidon/kubelet:v1.19.3
module.aws-ci1604702277-hz.null_resource.bootkube-start (remote-exec): Nov 06 22:42:23 ip-10-0-1-23 bootkube-start[3303]: Unable to find image 'quay.io/kinvolk/bootkube:v0.14.0-helm2-amd6' locally

https://yard.lokomotive-k8s.net/builds/42171

[iaguis]

lgtm