Public Repository

Last pushed: 7 months ago
Short Description
docker to demo mapr time series QSS
Full Description

Available metric to explore in opentsdb UI: ethylene, prediction, r1, r2, .... r16 (opentsdb metrics)
before run this image, make sure enough resources has been assigned to docker, I assigned 3 cores and 8GB memory

rm -rf /tmp/maprdemo
mkdir -p /tmp/maprdemo/hive /tmp/maprdemo/zkdata /tmp/maprdemo/pid /tmp/maprdemo/logs
chmod -R 777 /tmp/maprdemo/hive /tmp/maprdemo/zkdata /tmp/maprdemo/pid /tmp/maprdemo/logs

docker run -it --privileged -v /tmp/maprdemo/zkdata:/opt/mapr/zkdata -v /tmp/maprdemo/pid:/opt/mapr/pid -v /tmp/maprdemo/logs:/opt/mapr/logs -p 2222:22 -p 8443:8443 -p 4242:4242 -p 7077:7077 -p 8080:8080 mengdong/mapr-timeseries:latest

The docker will prepare itself for a while (about 5 minutes). When the spark job starts to generating logs, meaning everything is up and running. It will print out some example predictions and metrics every 5 seconds interval

In the browser, go to localhost:4242 to access the UI. In metric, type in "ethylene", add a new metric "prediction" and put the axis on the right. Then you can compare the prediction and actual time series data. The prediction is always 5 seconds ahead of time. Do auto-reload every 15 seconds. Don't specify stop time.

One advice is to run this as a demo is to begin running it before the presentation and pull the result after 5 minutes mark

Docker Pull Command
Owner
mengdong