Deploy Instructions

Overview

Describes how to set up and deploy the application to the WCI term-server.com server.

Details

Log into the server 10.230.174.16
1. You'll need to be able to sudo su - to root and ec2-tomcat.
Install and configure software as root (nginx, tomcat7 (and webapps), mysql-server-5.5 (and client), Java/jdk (including devel), maven, git, svn, etc.
1. Edit the /etc/init.d script to comment out the top part that forces running as root
2. Edit the /etc/init.d script to use this so the ec2-user can write the pid file:
  CATALINA_PID="/var/lib/tomcat7/$NAME.pid" # also change all instances of /var/run to /var/lib/tomcat7
3. Set permissions of the tomcat installation to the "ec2-tomcat"user
  # for each linked directory in /usr/share/tomcat7 d chown -R ec2-tomcat $dir chgrp -R ec2-user $dir Edit /etc/init.d/tomcat7 to use the following * TOMCAT_USER=ec2-tomcat * TOMCAT_GROUP=ec2-user Also edit /etc/tomcat7.conf * CATALINA_PID=/var/lib/tomcat7/tomcat7.pid * TOMCAT_USER=ec2-tomcat * TOMCAT_GROUP=ec2-user Make sure all the "chown" commands in the script change to group to $TOMCAT_USER
  NOTE: the tomcat user can start/stop the server with "/usr/sbin/tomcat7 start/stop".
  HOWEVER: if the process was started by root, it will likely have to be killed by root first.. we're working on this.
Tomcat must be configured to pass "run.config" environment variable to the application
1. One way to do this is to invoke the java command with -Drun.config=/home/ihtsdo/data/config.properties
  (e.g. edit this in /usr/share/tomcat7/conf/tomcat7.conf and add to JAVA_OPTS)
2. Also add -Xmx2500M to JAVA_OPTS in the tomcat startup script.
3. NOTE: development environments must also be configured as such
Tomcat must be configured to support POST requests of size > 2MB
1. Edit server.xml and add maxPostSize="50000000" to each <Connector> element.
Tomcat must be configured to support following of links. Make sure any Context tags in the configuration contain allowLinking="true".
1. Edit context.xml to add this parameter to the Context tag.

To enable tomcat manager, add this to tomcat-users.xml file.

<role rolename="manager-gui" />
<user name="tomcat" password="********" roles="manager-gui" />
<role rolename="protected" />

Make sure you have MAVEN_OPTS set
1. e.g. -XX:MaxPermSize=512m -Xmx3072m
Nginx Configuration
1. Install the SSL keys as root in the appropriate places (if SSH needed)
2. Add a file "snomed.term-server.com.conf" to /etc/nginx/conf.d
3. Add a Valve to tomcat to handle proxy headers.
```
        <Valve className="org.apache.catalina.valves.RemoteIpValve"
               remoteIpHeader="x-forwarded-for"
               remoteIpProxiesHeader="x-forwarded-by"
               protocolHeader="x-forwarded-proto"/>
```
4. Edit /etc/nginx/nginx.conf and add "client_max_body_size 50M;" to the "http" context.
5. /etc/nginx/nginx.conf contains the master configuration - see for details
6. For more info click here
As the MySQL root user, create an OTF user with the mysql client

CREATE USER 'otf'@'localhost' IDENTIFIED BY 'otfpwd';
GRANT ALL privileges ON *.* TO 'otf'@'localhost';

As the MySQL root user, create a database in the mysql client

CREATE DATABASE tsdb
  DEFAULT CHARACTER SET utf8
  DEFAULT COLLATE utf8_general_ci;
CREATE DATABASE tsdbtest
  DEFAULT CHARACTER SET utf8
  DEFAULT COLLATE utf8_general_ci;
GRANT ALL ON tsdb.* TO 'otf';
GRANT ALL ON tsdbtest.* TO 'otf';

Get code

cd
mkdir code
cd code
git clone https://github.com/WestCoastInformatics/SNOMED-Terminology-Server.git .
 
# switch to dev branch - for UAT
git checkout develop
 
# switch to dev branch - for PROD
git checkout master
 
mvn -DskipTests=true clean install

The sample data lives in the "config" project. Install in the ~/data directory

cd
mkdir data
cp -r code/config/src/main/resources/* data

Prepare a config file

cd
mkdir config
cp -r code/config/src/main/resources/config.properties.prod config/config.properties
# edit as needed

Prepare the tomcat indexes location

# To clean up the indexes do this (as root)
/bin/rm -rf /var/lib/tomcat7/indexes
mkdir /var/lib/tomcat7/indexes
mkdir /var/lib/tomcat7/indexes/lucene
mkdir /var/lib/tomcat7/indexes/lucene/indexes
chmod -R ga+rwx /var/lib/tomcat7/indexes
chown -R ec2-tomcat:ec2-user /var/lib/tomcat7/indexes
 
# Ensure the config.properties file used to run the application has this setting
hibernate.search.default.indexBase=/var/lib/tomcat7/indexes/lucene/indexes

Create and Load the data

cd ~/code/admin/loader/scripts
load.sh >&! load.log

Stop tomcat, clear cache,and remove the currently deployed .war files (and associated folders).

# Stop the tomcat7 server
service tomcat7 stop
/bin/rm -rf /var/lib/tomcat7/work/Catalina/localhost/mapping-rest
/bin/rm -rf /var/lib/tomcat7/webapps/term-server-rest
/bin/rm -rf /var/lib/tomcat7/webapps/term-server-rest.war

Update the database (this needs to go into the reload page):

mvn clean install -Drun.config=/home/ihtsdo/config/config.properties -Dhibernate.hbm2ddl.auto=update
Update file modification date to prevent caching issues.

cd ~/code
set date = `date +%Y%m%d%H%M`
touch -t $date `find . -name "*"`
Build Rest and deploy war

cd ~/code/rest
mvn -Drun.config=$OTF_MAPPING_CONFIG clean install
/bin/cp -f target/mapping-rest*war /var/lib/tomcat7/webapps/mapping-rest.war
Build Webapp war

cd ~/code/webapp
mvn -Drun.config=$OTF_MAPPING_CONFIG clean install
/bin/cp -f target/mapping-webapp*war /var/lib/tomcat7/webapps/ROOT.war
Start service and link "doc" directory.

service tomcat7 start

# wait for deploy to create the ROOT directory
sleep 5
# link doc directory
cd /var/lib/tomcat7/webapps/ROOT
ln -s /home/ihtsdo/data/doc
NOW, if deploying to PROD, create a github pull request.

Wire tomcat to redirect to the home URL. In the webapp root find the ROOT application and replace the index.html page. e.g.
NOTE: DO NOT do this step anymore as the application is now deployed as the root war

<?xml version="1.0" encoding="ISO-8859-1"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
   "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
    <meta http-equiv="refresh" content="0; url=/mapping-webapp/app/index.html#/" />
    <title>IHTSDO Mapping Tool - redirect</title>
</head>
<Body>
</body>
</html>

Complete Data Load

Get all data
- http://mapping.snomedtools.org/data/dev.zip (for dev data) - old
- http://mapping.snomedtools.org/data/prod.zip (for prod data) - old

Getting Backups from PROD to UAT

A daily backup tgz file exists on the prod server at

/var/backup/data.tgz

Alternatively, you could log into the uat and use this command

sudo aws s3 ls mapping.backup.ihtsdo

This will produce a list such as the following (just run):

2014-09-28 01:02:47 343656836 dbbackup_mappingservicedb_201409280100.bak.gz
2014-09-29 01:02:48 343750530 dbbackup_mappingservicedb_201409290100.bak.gz
2014-09-30 01:02:49 343867437 dbbackup_mappingservicedb_201409300100.bak.gz
2014-10-01 01:02:46 343892129 dbbackup_mappingservicedb_201410010100.bak.gz
2014-10-02 01:02:46 349749641 dbbackup_mappingservicedb_201410020100.bak.gz
2014-10-03 01:04:15 349779484 dbbackup_mappingservicedb_201410030100.bak.gz

Choose the backup you want to copy (it keeps the most recent 6 days) and do the following (this example taking the most recent backup):

sudo aws s3 cp s3://mapping.backup.ihtsdo/dbbackup_mappingservicedb_201410030100.bak.gz ~/

This will copy to ~/ (or wherever you want by changing that location at the end of the command). Then gunzip the file.