This documentation relates to Confluence 3.1.x
If you are using an earlier version, please view the previous versions of the Confluence documentation and select the relevant version.

Confluence Installation Guide

Welcome to the Confluence Installation Guide!

What is Confluence?

Confluence is an enterprise wiki that makes it easy for your team to collaborate and share knowledge. A wiki is a web application that lets you edit web pages easily and immediately. No waiting, just click, type, and click again.

Want one? Let's go!

Before you Start

Please check the following points:

  1. Make sure that your system meets the minimum requirements to run Confluence:
    • If you are installing Confluence for evaluation purposes, it should be pretty easy. You can use the Confluence Installer for Windows or Mac, or the zip archives. You will need a web browser — we recommend Firefox or Internet Explorer.
    • For production installations, use the zip archives. Please read the detailed system requirements.
  2. Please verify that this version of the Confluence documentation matches that of the Confluence version you are installing. The Confluence documentation version you are currently viewing is indicated toward the top of the page tree on the left or in the 'breadcrumb trail' in the top banner of this page. If you need to access a different version of the Confluence documentation, use the control at the top of the page tree on the left or you can access it from the documentation home page.

Installing and Setting Up Confluence

Two phases: Installation and Setup

There are two phases to the Confluence installation and setup procedure. The instructions on this and following pages will lead you through both phases. Overview:
  • Install Confluence onto your computer. We provide detailed instructions for installing the Standalone distribution and the EAR/WAR distribution.
  • Provide initial setup information. The Setup Wizard will prompt you for the information needed to get you up and running.

Choose the type of Confluence installation you'd like from the table below, and follow the link(s) to the remaining installation instructions. When you have finished the installation phase, you will be prompted to start the setup phase.

Installation Type Description
Evaluation installer for Windows
or
Evaluation Installer for Mac OS X
These are the simplest options for installing and evaluating Confluence. Choose these options if you wish to evaluate Confluence on a Windows- or a Mac OS X-based system.
These installers are specialised versions of the Standalone distribution (below) that should only be used to evaluate Confluence.

Standalone distribution The 'Standalone' distribution provides the simplest method of installing Confluence for long term or production use. Choose this option if any of the following is true:
  • You are evaluating Confluence and may likely use this installation for long term or production purposes.
  • You want to run the Confluence application on the Apache Tomcat application server bundled with this distribution.
  • You are not sure what you want, except to get Confluence up and running with minimum fuss.
EAR/WAR distribution This distribution allows you to deploy Confluence onto your own existing application server, instead of the Apache Tomcat server bundled with the Standalone distribution.
Confluence Clusters Please read the Confluence Clustering Overview and the Cluster Checklist before you consider installing Confluence in a cluster.
Upgrading Confluence Choose this option if you want to upgrade an existing Confluence installation to a new release of Confluence.
RELATED TOPICS

Upgrading Confluence
System Requirements

Labels

installation installation Delete
database database Delete
requirements requirements Delete
confluence confluence Delete
system system Delete
install install Delete
website-link website-link Delete
sizing-guide sizing-guide Delete
Enter labels to add to this page:
Please wait 
Looking for a label? Just start typing.
  1. Oct 08

    Anonymous says:

    Cleaning out the System Tasks Status Log This article was formerly published as...

    Cleaning out the System Tasks Status Log

    This article was formerly published as KB-article 181-2775 in Blackboard's Parature knowledgebase.

    Issue

    The System Tasks Status Log needs regular cleaning.

    Symptoms

    There are many entries listed in the System Tasks Status log, viewed from the system administrator panel. Deleting them manually takes far too long, and entries in "error" status cannot be deleted at all.
    In the performance report on Unix platforms, the "get_next_task" procedure shows up as using significant database resources.

    In Blackboard release 9, a large amount of additional entries with task_type='A' is being produced in this table. This necessitates very frequent cleanup (ideally daily, especially if daily enrollment snapshots are being run). Please see the additional information about [the NotificationRemoveStaleData task].
    Also note that due to the large amount of extra entries, the previously suggested index on the STATUS column is no longer sufficient, and a multi-column index on (STATUS,TASK_TYPE) is suggested instead. See below.

    Analysis

    These entries are not needed over time, as they are only useful to determine if a recent course copy or archive restore was executed successfully. If the status is "error", this serves as an alert that you should redo that course copy or additional operation. Having many of these entries leads to performance problems on the system, since the table is checked frequently for any tasks to be executed, and full table scans will be executed every time on the QUEUED_TASKS table.

    Adding an index for performance

    Even if for some reason you don't want to clean out these older records you will still want to avoid the resulting performance hit – and note that even with daily cleaning of the table the number of records in release 9 warrants adding this index. For this purpose every client should add an index on QUEUED_TASKS(STATUS,TASK_TYPE) (and Oracle users need to collect histograms for this column). This will avoid the full table scans.

    Oracle instructions

    Execute these two commands while logged into the BB_BB60 schema via sqlplus:

    CREATE INDEX QUEUED_TASKS_ST ON QUEUED_TASKS(STATUS,TASK_TYPE) TABLESPACE BB_BB60_INDX;
    EXEC DBMS_STATS.GATHER_TABLE_STATS(ownname=>'BB_BB60', tabname=>'QUEUED_TASKS',
    cascade=>TRUE, method_opt=>'FOR COLUMNS STATUS,TASK_TYPE SIZE AUTO');
    
    SQLserver instructions
    • In SQLserver Management Studio, open the BB_BB60 database and select the QUEUED_TASKS table.
    • In the Object Explorer window at right, click on Indexes (or expand QUEUED_TASKS at left and click Indexes below it).
    • Right-click in the background and choose "New Index".
    • Name the index "QUEUED_TASKS_ST".
    • Click on "Add" to specify columns to be indexed.
    • Select the STATUS and TASK_TYPE columns.
    • In the result, TASK_TYPE will be listed above STATUS. Use the MOVE buttons to move STATUS to the top.
    • OK/close the new Index dialog. The index will be created right away.

    Alternative steps:

    • Open the table in design mode, then click on the "Indexes/Keys" button on the toolbar (near right end, it shows a key-and-list icon).
    • click on "Add", then for the newly added index find "General > Columns" and click on the right field (where it currently says "pk1"). Click the ellipsis button ("…") that now shows up at the right.
    • Select the STATUS column in the drop-down list. Leave order as ascending.
    • Click in the empty field below "status" and select TASK_TYPE from the drop-down. Again leave order as ascending. Click "OK".
    • Under "Identity > name", enter QUEUED_TASKS_ST (instead of the suggested "IX_QUEUED_TASKS").
    • Under "Table Designer > Recompute Statistics" select "Yes". Click "Close".
    • Close the table design view and choose to save the design changes (or hit Ctrl-S before closing the view).

    Resolution

    This logging information is kept in the database. You can run the following query which will delete all completed tasks older than 7 days, as well as those in "Error" status, which are not deletable from the GUI. Connect to the BB_BB60 database and run the appropriate version for your database type:

    Oracle
    DELETE FROM queued_tasks WHERE status IN ('C','E') AND end_date < sysdate-7;
    COMMIT;
    
    SQL Server
    DELETE FROM queued_tasks WHERE status IN ('C','E') AND end_date < getdate()-7
    

    You may also want to delete more recent tasks, and besides tasks with 'C' (completed) or 'E' (error, not restartable) status you may want to also delete tasks in 'I' status (incomplete, can be restarted) as well - this is up to the user.

    For a very quick and complete cleanup, check that no tasks are in 'A' (assigned to be run), 'R' (running), or 'W' (waiting) status, then run TRUNCATE TABLE QUEUED_TASKS. The following PL/SQL code does this for you in Oracle:

    DECLARE
    v_count NUMBER;
    BEGIN
    SELECT count(*) INTO v_count FROM QUEUED_TASKS WHERE status IN ('A','R','W');
    IF v_count=0 THEN
    EXECUTE IMMEDIATE 'TRUNCATE TABLE QUEUED_TASKS';
    END IF;
    END;
    /
    
    Also note that occasionally a course copy task may have gotten stuck and may prevent other such tasks from completing - if that happens, try deleting the offending task (first not-completed task) listed in the GUI from the database side and things should work again after a server restart. The task should be identified by a status of 'R' (running) and a task_type of 'C' (course copy).

    Scheduling regular cleanup

    You may want to schedule the above as a regular cleanup task in your SQLserver maintenance plan, or as an Oracle job via some code like this (of course you could schedule this to run less frequently, or to keep more data, or run at a different time/day):

    DECLARE
    JobNo dba_jobs.job%TYPE;
    BEGIN
    --schedule cleanup of QUEUED_TASKS each Sunday morning at 4:30am:
    DBMS_JOB.SUBMIT(JobNo, 'DELETE FROM queued_tasks WHERE status IN (''C'',''E'') AND end_date < sysdate - 7;',
    trunc(sysdate) + 4/24 + 1/48 + 8 - to_char(sysdate,'D'),
    'trunc(sysdate) + 4/24 + 1/48 + 7');
    END;
    /
    
    Unknown macro: {show-to}
    This is former Swiki article [Support/445]
  2. Oct 08

    Anonymous says:

    test

Add Comment