Deploy Instructions

Overview

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

Details

1. Log into the server 10.230.174.16
   
   1. You'll need to be able to sudo su - to root and ec2-tomcat.
      
      
2. Install and configure software as root (nginx,tomcat7, mysql-server-5.5, 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. 

       

3. 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
      
      
4. Tomcat must be configured to support POST requests of size > 2MB
   1. Edit server.xml and add  maxPostSize="50000000" to each <Connector> element.
      
      
5. 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.
      
      
6. Make sure you have MAVEN_OPTS set 
   
   1. e.g. -XX:MaxPermSize=512m -Xmx3072m
7. 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
       

8. 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';

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

    

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

10. Get code

     

    cd mkdir code cd code git clone https://github.com/WestCoastInformatics/OTF-Mapping-Service.git .   # switch to dev branch - for UAT git checkout develop   # switch to dev branch - for PROD git checkout master   mvn -DskipTests=false clean install

11. Get data.

     

    cd mkdir data git clone ssh://csfe.aceworkspace.net:29418/mapping data mkdir data/doc mkdir data/doc/archive chmod -R ga+rwx data/doc

12. 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 ihtsdo /var/lib/tomcat7/indexes chgrp -R ihtsdo /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

     

    
    
    

13. Create and Load the data

     

    cd ~/code/admin/loader/scripts load_prod.csh >&! load_prod.log

     

    Note also that TreePositions will be automatically calculated when loading a terminology.  To recompute SNOMED tree positions without reloading terminology, use:

     

    cd ~/code/admin/loader mvn -Drun.config=prod -P SNOMEDCT-treepos

14. 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/mapping-rest /bin/rm -rf /var/lib/tomcat7/webapps/ROOT /bin/rm -rf /var/lib/tomcat7/webapps/mapping-rest.war /bin/rm -rf /var/lib/tomcat7/webapps/ROOT.war

15. 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

16. Update file modification date to prevent caching issues.

     

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

17. 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

18. 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

19. 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

20. NOW, if deploying to PROD, create a github pull request.
    
    

21. 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

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):

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