Apply all necessary steps from R Consulting Appliances Installation, Initial configuration checklist first.

HLS Accelerator (hac) Appliance configuration checklist

  1. Initialize cache volume, add to /etc/rc.conf.d/datamount:

    datamount_enable="YES"
datamount_vol1="gpt/vol1"
datamount_wwwcache="vol1"

  2. Cache configuration in /usr/local/etc/nginx/local.hac-cache.conf:

    # Fields to customize: cache size of 10M fits about 75k URLs,
# inactive depends on site configuration, max_size on free space
proxy_cache_path /srv/wwwcache
                 levels=1:2 keys_zone=upstream.cache:32M
                 inactive=48h max_size=320G
                 use_temp_path=off;

    Change inactive and max_size and refer to nginx documentation if must.

  3. Configure upstream in /usr/local/etc/nginx/local.http.upstreams.conf:

    upstream us1 {
    server upstream:80;
    keepalive 4;
}

    us1 is upstream id, as referenced in the upstreams.lua script (see below). upstream is the host name of source server. It could be simply left as upstream, as long as respective alias is added to /etc/hosts. One file can list several upstreams.

  4. Create /etc/rcons/hac/upstreams.lua:

    return {
    ["1.2.3.4"] = "us1", -- <-- storage IP
    ["1.2.3.5"] = "us1", -- <-- my IP
    ["acme.com"] = "us1" -- <-- my host name (for SSL)
};

  5. (Optionally) enable CORS, required for HLS playback on some devices; create /usr/local/etc/nginx/local.conf.cors.conf:

    map $http_origin $cors_origin {
    default "";
    "~^https?://tv.acme.com(:[0-9]+)?$" "$http_origin";
}

    and replace tv.acme.com with mobile TV service address’ domain.

Configuring cascading HACs

  1. Add to /usr/local/etc/nginx/local.http.upstreams.conf on the edge HAC:

    upstream us1 {
    server upstream:80;
    keepalive 4;
}

    where us1 and upstream have same meaning as when no cascading is in use, however us1 is passed as Host: HTTP request header when edge is addressing the caching HAC.

  2. Add to /usr/local/etc/nginx/local.http.revproxy.conf on the caching HAC:

    server {
    listen 80;
    server_name us1;
    allow 192.168.0.0/16;
    deny all;

    location / {
        proxy_http_version 1.1;
        proxy_pass http://us2;
        proxy_cache upstream.cache;
        proxy_cache_valid 200 24h;
        proxy_set_header User-Agent nginx/$nginx_version;
        proxy_set_header RcHlsCdn-Accel-Id $hostname;
        proxy_set_header RcHlsCdn-Accel-Host $http_rchlscdn_accel_host;
        proxy_set_header RcHlsCdn-Accel-Peer $http_rchlscdn_accel_peer;
        proxy_set_header RcHlsCdn-Accel-UA $http_rchlscdn_accel_ua;
        proxy_set_header Connection "";
    }
}

    where (server name) us1 is the upstream id under which edge HAC references caching HAC in the previous point, and us2 is the upstream id for caching HAC’s upstream. 192.168.0.0/16 is IP subnet edge HACs will accessing caching HAC from.

Resources