ecFlow's documentation is now on readthedocs!
The following diagram shows the typical status changes for a Task.
In addition, the status of a 
Let us have a look at  ecflow_ui, this provides the GUI for ecFlow.
On the UNIX prompt type ecflow_ui

Once the program is up:

  • Select “Servers” menu. 

  • Next click on the "Manage Servers..." menu

  • "Add server" enter details of your ECF server.

You should also change the preferences to "Administrator" mode, which shows an enhanced set of menu options.

  • In ecflow_ui: Select "Tools Menu:Preferences"
  • Select Menus and change "menu mode" to "Administrator".

Choose a name for your ECF server and add the Host name and port number you noted down earlier. Make sure 'Add server to current view' is clicked on the dialog.

You should get something similar to the following display :

The yellow boxes are called nodes. With the left mouse button click on the black triangle. This will expand the node. Repeat for node called 'test'.

S - stands for the suite, F stands for Family, and T - stands for Task

Now select the View menu and then "Add info panel". This panel can be re-arranged on the window.

You can now view the ecf script, the job file and the output of the task t1.
For that click on the node t1, then press on the script tab.

To see the ECF job or the output simply click on the tabs.
If you want to see the output in a different window, select the node t1, then with the right mouse button, select the "Output..." menu option.

If we want to rerun the suite, move the mouse pointer over test and click on the right mouse button, and choose requeue.

You will notice that the node‘s change colour.
The colours reflect the status of the nodes.
There can be a long delay between the job changing status and that change being reflected in the ecflow_ui window (set to 60s as default).
To update the current state more quickly in that window, click the refresh (or F5).

We will now add a table view. This allows us to see additional attributes when a node is selected. From the View menu, select "Add table view"

You will notice that selection in the tree view, selects the corresponding node in the table view. 

Task nodes contain a flag that can reflect the reason for its being aborted, complete, or submitted.

Here the associated icons that can be displayed in ecflow_ui

What to do

  1. Explore ecflow_ui menus and windows
  2. There can be a long delay(60s) between the job changing status and that change being reflected in the ecflow_ui window.
    To update the current state more quickly in that window, click on the green icon(top left-hand side), on the toolbar. (or F5)


  1. following the arising of few questions, here the reminder about another set of icons

    a child command is trying to communicate with the ecFlow server, with an incorrect password.
    use menu Special->FreePassword or edit ECF_PASS variable to FREE to allow connection.
    ServerNode->Zombies menu will also open a dedicated panel

    zombie directive was added to the definition file for automatic zombie handling

    this can be displayed setting Show->Icons->Complete. It reflects that whether
    defstatus complete is defined for this node (static) or a
    complete attribute (dynamic) is attached to the node, with a valid expression, that causes this node complete.
    We can then distinguish the tasks completed thanks to ecflow_client --complete called from those which may not have to run.

    Show->Icons->Complete must be unselected when we want all complete nodes to disappear from the display

    this icon reflects that the active job is blocked (livelock) calling ecflow_client --wait <trigger_expression>
    with the trigger expression not yet true

    late icon may appear for a task where late attribute was attached, reporting latency for submission, runtime or completion; it may then cause a popping window. On the server node, it report latency writing checkpointfile or a jobfile.A menu->Special is available to clear this flag.

  2. popping window can be shown, when a task is restarted , aborted ,  or task is tagged with zombie or late flag.

    Global setting is Tools→Preferences  to activate or not the popping window.

    This choice can be overwritten with ServerNode→Info Panel->Settings(tab) (local) setting.

    Flags can be changed with alter command for tasks and family, as it can be used to tag/find a set of nodes ; however, only tasks shall cause the related popping window to be raised.

    The window manager may be configured to prevent "focus stealing" when necessary.