Updated notes and comments on server configuration

config/environment.sh

@@ -1,3 +1,5 @@
+# Copy this file to /etc/blot/environment.sh and fill it in
+
 # Flags
 export BLOT_PRODUCTION=true
 export BLOT_CACHE=true

config/guides/production-environment-setup.txt

@@ -11,6 +11,36 @@ Then make sure you make all the contents of scripts/production executable:
 chmod +x scripts/production/start_blot.sh
 ```
 
+AWS
+---
+
+At the moment I have moved Blot to an I3 instance which has 16GB of RAM, a 500GB SSD hardrive for temporary data.
+
+The 500GB harddrive is what Amazon calls an 'instance store'. It should not be depended on. When the instance reboots the data on it is lost, I believe.
+
+I think it might be a good idea to use it as a cache. I should also move my EBS data to a cheaper (non-ssd)drive. Then I can use the SSD drive that comes with the i3 instance.
+
+The I3 instance offers volatile storage.
+
+---
+
+3/3/2018 I have since set up the instance store as a cache.
+
+I followed these steps to set up the disk to use as a cache:
+
+https://richardstechnotes.wordpress.com/2015/12/18/setting-up-an-nvme-ssd-on-ubuntu-14-04-lts/
+
+More instructions here:
+
+https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/add-instance-store-volumes.html
+
+I don't really care about write speed (as opposed to read speed) but here is a guide for optimizing write performance on instance stores:
+
+https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/disk-performance.html
+
+
+
+
 Production machine configuration (Linux)
 ----------------------------------------
 
@@ -20,6 +50,7 @@ sysctl vm.overcommit_memory=1
 sudo bash -c "echo never > /sys/kernel/mm/transparent_hugepage/enabled"
 dd if=/dev/zero of=/swapfile1 bs=1024 count=4194304
 
+
 Upstart
 -------
 

config/monit.rc

@@ -1,6 +1,7 @@
 # Monit exists to restart the node.js server or
 # NGINX if either become unresponsive. Upstart
 # only restart these processes if they exit.
+
 set logfile /var/www/blot/logs/monit.log
 
 # The route which responds is in app/index.js

config/readme.txt

@@ -3,23 +3,24 @@ Config
 
 This directory contains the scripts and configuration code to run NGINX, Redis and Monit, which combine with the node js server to produce Blot. When setting up a new production or development machine, follow the respective guides in the guides directory. There are a number of scripts for updating Blot's configuration on the server, which are stored in scripts/production:
 
-Upstart: reloading the scripts
-sudo ./scripts/production/update_upstart.sh
+Upstart: reloads the scripts to start Blot, nginx, redis and monit
+$ ./scripts/production/update_upstart.sh
 
-Monit: reloading configuration
-sudo ./scripts/production/reload_monit_rc.sh
+Monit: reloads its configuration
+$ ./scripts/production/reload_monit_rc.sh
 
-NGINX: reloading configuration with zero downtime
-sudo ./scripts/production/nginx_reload_configuration.sh
+NGINX: reloads its configuration with zero downtime
+$ ./scripts/production/nginx_reload_configuration.sh
 
 
 Logging
 -------
 
-Check logs/sys.log for upstart related logging
-Check logs/monit.log for monit related logging
-Check logs/app.log for the node.js server's logs
-Check logs/nginx.log for nginx's access log
+Check ../logs/sys.log for upstart related logging
+Check ../logs/monit.log for monit related logging
+Check ../logs/app.log for the node.js server's logs
+Check ../logs/nginx.log for nginx's access log
+
 
 Upstart
 -------