adler

# I was able to capture all the http traffic on port 9740
$ sudo tcpdump -s0 -w madler-posts.pcap port 9740

# Here we can the health checks that GET /version
$ tcpdump -r madler-posts.pcap -A port 9740 | grep GET | head
reading from file madler-posts.pcap, link-type EN10MB (Ethernet)
...w7Q.{GET /version HTTP/1.1
....7Q..GET /version HTTP/1.1
....7Q..GET /version HTTP/1.1
.5.B7Q..GET /version HTTP/1.1

adler

#!/bin/bash

# runs a command in the same network namespaces as a process pid

set -e

function usage {
    test -n "$*" & echo "$*"
    echo "usage: $0 PID COMMAND [ARG...]" >&2
    exit 1

adler

    "appName" : {
      "Type" : "AWS::ElasticLoadBalancing::LoadBalancer",
      "Properties" : {
        "Subnets" : [{ "Ref" : "SubnetA" }, { "Ref" : "SubnetB" }, { "Ref" : "SubnetC" }],
        "SecurityGroups" : [ { "Ref" : "lbSG" }  ],
        "CrossZone" : true,
        "Policies": [ {
          "PolicyName" : "TLSOnlyPolicy",
          "PolicyType" : "SSLNegotiationPolicyType",
          "Attributes" : [

adler

#  null_backend is always unhealthy
backend null_backend {
    .host = "localhost";
    .port = "39997";
    .probe = {
        .url = "/";
        .timeout = 5ms;
        .interval = 60s;
        .window = 2;
        .threshold = 1;

adler

# get half of number of bytes of total memory
SHMMAX=$(free | perl -ane 'printf "%1.0f\n", ($F[1] * 512 ) if m/Mem:/')
grep -q kernel.shmmax /etc/sysctl.conf || echo "kernel.shmmax =  $SHMMAX" | sudo tee -a /etc/sysctl.conf
sudo sysctl -p