The platform introduced in version 3 includes important changes that affect the way you install and administer the application. This topic provides an overview of those changes, the platform, and the tools that come with it.
Here's a high-level list of the things that are new or changed with the platform's introduction:
You might be interested in other documentation related to the platform. The following topics include references and guides for making the most of the platform and applications installed on it.
| Document | Description |
|---|---|
| Platform Run Book | Basic system commands for managing the platform. If you need to handle something quickly, start here. |
| Operations Cookbook | Sample configurations and script examples for long-term operations. |
| Application Management Commands | A reference for commands you can use to maintain your managed instance. |
| System Requirements | System components and technologies supported in this release. |
| Installation Guide | Step by step instructions for installing the platform using RPM. |
The Jive SBS platform consists of several high-level components that work together to provide overall system functionality. The following sections provide an overview of these components and their role in the architecture. By default, the RPM will install the following components to start and stop in the correct order during system startup and shutdown.
The platform contains a customized version of the Apache Foundation’s Apache HTTPD web server software. Among other things, this software is used to process HTTP and HTTPS-encrypted requests between the platform and end users. The Apache HTTPD server uses the JK protocol to communicate with the back-end application server described in the next section.
The Apache Tomcat application server is used to host back-end application and business logic as well as to generate dynamic HTTP traffic served by the Apache HTTPD Server. The application server is also responsible for processing email delivered by the system, and for scheduling background tasks such as search index management and integration with third-party applications and systems.
The PostgreSQL database server is an RDBMS (Relational Database Management Server) service that is hosted internally within the platform. You can use this service or a PostgreSQL or Oracle system of your own.
By default, the Jive SBS RPM will attempt to install a local system user named “jive” belonging to group “jive” with a home directory of “/usr/local/jive”. If a user named jive already exists on the system where the RPM installation is being performed, the RPM will use the existing user. Most binaries, scripts and configurations used by the platform are owned by the jive user with a small number of files owned by root.
The Jive SBS platform installed by the RPM is designed to manage one or more logical “application” servers, each serving a variety of functional concerns.
The master application template is the core set of Jive software used to create all subsequent instances. By convention, the files in the master template are stored in the jive user’s “applications” directory located at:
/usr/local/jive/applications/template
Upon installation, and at the request of system administrators, the platform will create instances of the master application template, each with a unique name. Each instance is managed, configured and runs independently of other applications. By design, an instance of an application runs in a dedicated operating system process, providing separation between applications such that one application instance cannot directly exhaust the resources of another in a way that cannot be stopped with standard operating system commands (such as kill).
An application instance is considered managed if it is created, updated, and its lifecycle controlled by the platform tooling (appadd, appstop, appstart, etc.). Managed applications are automatically upgraded when the RPM is upgraded.
In some circumstances, you might want to create applications that are not fully managed. For example, applications created with the “—no-link” options to the appadd tool will not be directly upgraded by platform upgrades. Non-managed applications are not directly upgraded by the Jive tooling and must be upgraded manually by a system administrator.
Ultimately, most Jive SBS tools delegate to standard Linux systems management tools such as bash scripts, the “ps” command, and so on. It is possible to monitor Jive-managed components using standard tools such as ps and top as described in the Platform Run Book.
When using the local database (installed as part of Jive SBS) as the system of record for a deployment, the platform will make an attempt to tune and back up the underlying database. While sufficient for smaller databases, larger databases will have storage and backup considerations unique to their individual deployments; you'll likely want to alter the platform defaults.
Upon installation, the Jive SBS platform database is scheduled for daily full backups via an entry in the cron.daily section of the system cron task scheduler. Specifically, the installation process creates a symlink that maps /usr/local/jive/bin/dbmaint to /etc/cron.daily/jive-dbmaint. This script performs a full analysis and backup of the contents of the local database. Backups are stored to /usr/local/jive/var/data/backup/postgres/full and are stamped with the date that the backup began. Logs for the backup and maintenance are stored in the standard platform location at /usr/local/jive/var/logs, in files dbmaint.log, dbback.log and dbanalyze.log.
In addition to full backups, the platform captures PostgreSQL WAL (Write-Ahead Log) segments to /usr/local/jive/var/data/backup/postgres/wal. For more information about WAL segments, see the PostgreSQL documentation.
The platform backup infrastructure will purge full backups that are 30 days or more old. The platform will not purge WAL segment backups described in the previous section. As a general rule, you can safely remove WAL segments older than the most recent full backup. You should keep WAL segments since the previous full backup so that you have it for a partial recovery of the database in the event of corruption.
System administrators should evaluate the storage capacity of their systems and the size consumption to arrive at the best strategy for purging backup files.
The platform manages one or more instances of the Jive SBS software. Each instance represents a dedicated Apache Tomcat application server that ultimately runs as a dedicated server process, independent of other application servers on the host machine. By convention, each application instance is represented by a directory entry beneath the “applications” directory located in the jive user’s home directory (/usr/local/jive/). For example, an application named “biodome” will have a corresponding application directory located at:
/usr/local/jive/applications/biodome
Each application has a standard set of subdirectories beneath its top-level application directory:
The Jive SBS platform installation captures meaningful system data to various log files as well as providing application and system snapshot capabilities. With these, Jive support can more easily diagnose issues with the platform.
The primary mechanism for gathering support-related information on the platform is via the appsupport command, typically executed as the jive user.
When run, the appsupport command will combine multiple useful data points from the system and application logs, then aggregate the data to the console’s standard output stream. Alternatively, you can use the “-o” flag, along with the path to a file where the aggregate system information should be appended (if the file does not exist, it will be created).
In some cases, Jive support may suggest that you capture application thread dumps from a running system. The platform includes the appsnap tool for this purpose. As with other commands, appsnap is commonly performed as the jive user.
The most common options used with the appsnap tool are the "—count" and "—interval" options. Count determines the number of samples taken. The interval option determines the number of seconds between successful samples. The tool writes snapshot output to the console’s standard output or appends it to the file given for the "-o" option.
If you're upgrading from a version prior to version 3, you might wonder about the jiveHome directory. The jiveHome directory was where you put (and found) the working files for your community. Caches, themes, plugins, logs -- that sort of thing.
| Resource | Location Prior to Version 3 | Location in the Platform |
|---|---|---|
| Application logs (including those for database backups) | <jiveHome>/logs | /usr/local/jive/var/logs |
| Installed plugins | <jiveHome>/plugins | /usr/local/jive/applications/<app_name>/home/plugins |
| User-created themes | <jiveHome>/themes | /usr/local/jive/applications/<app_name>/home/themes |
| Cached attachments | <jiveHome>/attachments | /usr/local/jive/applications/<app_name>/home/attachments |
| Cached images | <jiveHome>/images | /usr/local/jive/applications/<app_name>/home/images |
| Cached data | <jiveHome>/cache | /usr/local/jive/applications/<app_name>/home/caches |
| Resources directory | <jiveHome>/resources | /usr/local/jive/applications/<app_name>/home/resources |
| Local system database | <jiveHome>/database | /usr/local/jive/var/work/sbs/data/postgres-8.3 |
| jive_startup.xml file | <jiveHome>/jive_startup.xml | /usr/local/jive/applications/<app_name>/home/jive_startup.xml |
