installation_automatic.md

DS Q-P + OPAL installation guide

Install your poll thread, opal server and poll monitor server on a server

1 - prepare for installation

1. connect to your new server via ssh

2. change to sudo user: sudo -s

3. add the ssh user to access the prive ds repos

create a .ssh dir: mkdir ~/.ssh upload the id_rsa_ds file to your .ssh dir (get the private key as part of your installation package)

upload the prepare_repository_and_server.sh from your installation package to your server to the ~ directory

and change the rights to execute chmod +x ~/prepare_repository_and_server.sh and execute the prepare_repository_and_server.sh

2 install opal with R server and test data + the poll mechanism with gui

please note that this script will download ca. 2GB of data

open the opal_poll.config file and change the configuration information to your server requirements,

then execute ./install_prod_opal_poll.sh in this repo

in your browser open the poll mechanism monitor: <server ip of current server>:80/poll-monitor you should now see the poll monitor gui, which lets you control the poll thread

Install your queue server

1 - prepare for installation

1. connect to your new server via ssh

2. change to sudo user: sudo -s

3. add the ssh user to access the prive ds repos

create a .ssh dir: mkdir ~/.ssh upload the id_rsa_ds file to your .ssh dir (get the private key as part of your installation package) start the ssh agent and add the ssh key:

2 - start your queue server

upload the prepare_repository_and_server.sh from your installation package to your server to the ~ directory

and change the rights to execute chmod +x ~/prepare_repository_and_server.sh then execute the prepare_repository_and_server.sh

open the queue.config file and change the configuration information to your server requirements,

change your configurations in the install_prod_queue.sh file of this repository and then execute ./install_prod_queue.sh

working with the queue:

the queue is started in a docker container and can be managed with the following bash commands: first navigate to the queue repository cd ~/ds_deployment/ds_develop/ds_queue

starting the queue execute ./start.prod.sh

stopping the queue execute ./stop.prod.sh

getting the queue status execute docker exec queue_server_prod bash -c "cd /root/ds_queue && ./q_admin.sh status"

Test your installation

Your installation should now be complete and your queue server started. Go to your poll server installation and access its webservice to start the poll mechanism: in your browser: <poll_server_ip>/poll-monitor switch to the tab control and enter your queue server host and port and leave the opal server blank then click start poll server your poll server should no be running (see running label in user interface)

now go to your R datashield client server and execute the following commands in R, changing queue_server_host for your queue server host.

note: if you do not have a running R server see the section below Run R server locally

If you see this output:

Variables assigned:
datashield_opal--GESLACHT, GEWICHT, LENGTE, HEALTH17A1, HEALTH17B1, HEALTH17D1, DBPa, SMK11, SMK31, SMK4A1, SMK4A21

your Installation is working

# Load opal and datashield libraries
library(opal)
library(opaladmin)
library(dsBaseClient)
library(dsStatsClient)
library(dsGraphicsClient)
library(dsModellingClient)

# Login to VMs

# To understand why these variables are assigned this way, see the
# documentation for the datashield.login function (part of the opal
# package)

# login details
server <- c("datashield_opal")
# note the datashield_opal only works from inside this docker container
url <- c("https://queue_server_host:443")  # Ur quserver host or ip address here
# ^^^ Note this specifies the port number
user <- "administrator"
password <- "password"
table <- c("test.LifeLines")
# ^^^ note that this reflects the folder hierarchy that can be seen via the OPAL web interface

# Create a dataframe with all these details
logindata <- data.frame(server,url,user,password,table)

# Create an 'opals' object by passing the 'logindata' data frame to the
# datashield.login function
opals <- datashield.login(logins=logindata, assign = TRUE)

Run R server locally

As a first test you can start your test server on your poll server and connect to your queue.

For this navigate to the test repo:

cd ~/ds_deployment/ds_develop/ds_test and start the test docker docker-compose up -d then connect to your test docker container docker exec -it ds_test bash

start R R and then paste the R test above with the queue_server_host string changed to your queue server ip address