Add scripts to automate demo setup

Author: geoffxy

experiments/15-e2e-scenarios-v2/demo/full_workload_def.sh

@@ -27,7 +27,7 @@ trap "inner_cancel_experiment" TERM
 # Start with Aurora 2x db.t4g.medium, Redshift off, Athena unused.
 
 # Start with 4 analytical clients.
-start_repeating_olap_runner 4 15 5 $initial_queries "ra_4"
+start_repeating_olap_runner 4 25 5 $initial_queries "ra_4"
 rana_pid=$runner_pid
 sleep 2
 

experiments/19-demo2/README.md

@@ -0,0 +1,18 @@
+# Operating the BRAD Demo
+
+1. Start a new tmux session; you need 3 tmux panes
+2. Activate the virtualenv that has BRAD and its dependencies installed
+3. Run `./start_brad.sh` to start BRAD. The web interface should be accessible on port 7583
+4. In another tmux pane, run `./start_workload.sh` to start the workload
+5. BRAD relies on system metrics (retrieved from AWS) to do its planning. You
+   need to wait at least 3 minutes for the workload metrics to make it to AWS
+   before you can start the planner.
+6. To shut down the demo, shut down the workload runner first (Ctrl-C). Wait
+   until all runners have exited. Then shut down BRAD (Ctrl-C on the other pane).
+
+## Important files
+
+- `config/system_config_demo.yml`:
+  - BRAD configs for the demo (checked in)
+- `config/physical_config_100gb_demo.yml`:
+  - Physical configuration values; ensure the cluster IDs refer to actual AWS resources

experiments/19-demo2/start_brad.sh

@@ -0,0 +1,12 @@
+#! /bin/bash
+
+script_loc=$(cd $(dirname $0) && pwd -P)
+cd $script_loc
+
+pushd ../.. > /dev/null
+
+brad daemon \
+  --system-config config/system_config_demo.yml \
+  --physical-config config/physical_config_100gb_demo.yml \
+  --schema-name imdb_extended_100g \
+  --ui

experiments/19-demo2/start_workload.sh

@@ -0,0 +1,8 @@
+#! /bin/bash
+
+script_loc=$(cd $(dirname $0) && pwd -P)
+cd $script_loc
+
+pushd ../15-e2e-scenarios-v2/demo > /dev/null
+
+./02-start_workload.sh