DINA-Web Infrastructure

25 October 2015

Production server setup - Infrastructure stack

This setup uses the Ubuntu 14.04 LTS (64-bit) Linux OS.

Other important software components used here are:

  • DINA-Web modules - source code from repo at https://dina-web.github.io
  • Open source server software application stacks from https://bitnami.com
```console
 ____ ___ _   _    _
|  _ \_ _| \ | |  / \              +------------+
| | | | ||  \| | / _ \             |ssh gateway | devops1
| |_| | || |\  |/ ___ \            +------------+ devops2
|____/___|_|_\_/_/___\_\                 CC       devops3
\ \      / / ____| __ )                 /  \      devops4
 \ \ /\ / /|  _| |  _ \                /    \
  \ V  V / | |___| |_) |              /      \
   \_/\_/  |_____|____/              /        \
 https://dina-web.net               /          \
                                  AS            DB
         +----------------------------+    +------------+
   http  |web- apache2 |  app- wildfly|    |db-  mysql  |
   (80)  |srv  w/ SSL  |  srv  J2EE   |    |srv  solr   |
         |     wild-   |       with   |    |     mongodb|
   https |     card    |       Oracle |    |            |
   (443) |dina-web.net |       JDK v8 |    |            |
         +----------------------------+    +------------+
         subdomains      modules           datasets
           loans           /loans            mysql/dina_nrm
           inventory       /loans-admin      mysql/userdb
           loans-admin     /inventory        mysql/bot_test1
           issues          ...               mongodb/loans

```

Introduction to the infrastructure

In the above sketch, the CC server is an ssh gateway (Checkpoint Charlie) which provides access for devops users or sysadmins to the production environment that runs the DINA WEB system.

The application server AS contains a webserver which accepts web traffic (http, https) and reverse proxies it to Wildfly - a Java 2 EE application server, which runs the DINA WEB modules, which in turn gets their data from DB, the database server.

  • By using bitnami software stacks we get a “portable” setup so that backups of application and data files should be restorable and deployable even on another base OS. At least in theory. In practice the Ubuntu 14.04 LTS OS has been used.
  • A single OS user is used to run all the services. This avoids using the root user to install services etc.
  • Firewall redirection sends http and https traffic to services.
  • Startup scripts install the software to run as service on the OS, so it automatically starts up when the system is bounced.
  • Backup scripts are scheduled to run daily (for all data - “data”) and weekly (for all application files - “full”).

Currently operational DINA WEB services

See https://dina-web.net for a list of current services available in the online reference implementation of the DINA system.

Development infrastructure

Using vagrant and local VMs, locally. For international collaboration infrastructure, we are investigating needs for the continous integration servers and maintain a build server and test server on the Internet since for international collaboration purposes…

Open source dev collaboration tools

See the issue tracker using RedMine for various BIO projects…

We also use tools like:

  • Gitorious
  • Jenkins
  • Archiva (maven repo)
  • TRAC/RedMine
  • mailman
  • IRC

But may want to move further out into the clouds utilizing free hosted services (github instead of Gitorious etc, open an irc channel on Freenode? etc)