James Harr jamesharr

I'm having a problem getting FQDNs working, and I figured I'd post here to see if anyone could help.

The logs I'm seeing from hubble observe seem to fly directly in the face of the policy that's written.

The Kubernetes manifest is attached as dns-return-test.yml. The cluster is K3s (latest) running fairly stock. See below for details.

Things that seemed odd to me:

I had to explicitly allow return traffic from kube-dns to my pods, if that was not in the manifest, DNS would not work and hubble observe would show a weird ERROR. The specific area of the manifest is noted.
Disabling DNS inspection (noted in manifest) would cause DNS to work fine without the aforementioned rule. FQDNs wouldn't work, but DNS would.

	aws-token(){
	aws_token_sn=arn:aws:iam::REDACTED_ACCT_NUM:mfa/REDACTED_USER_PASSWD
	aws_token_duration=43200
	export AWS_ACCESS_KEY_ID=""
	export AWS_SECRET_ACCESS_KEY=""
	export AWS_SESSION_TOKEN=""

	read -p "MFA Code: " aws_token_code
	OUT="$(aws sts get-session-token \
	--serial-number $aws_token_sn \

	_start_tmux() {
	# Skip if we're not interactive
	[ -z "$PS1" ] \|\| return
	echo $- \| grep -qs i \|\| return

	# Skip if we're logged in via VSCode
	[ -z "$VSCODE_IPC_HOOK_CLI" ] \|\| return

	# Skip if we're not on an ssh connection
	[ -z "$SSH_CONNECTION" ] && return

	# ~/.ssh/config
	Host bingbingbong
	HostName 192.0.2.1

	PermitLocalCommand yes
	LocalCommand say '[[volm 0.5]]' bing bing bong &