Manage Jobs, views and limits
Open Run → Job Manager from the Library. The list shows the executions; select a row to see the log snapshot in the bottom panel. Select all selects the panel text; Copy details copies the details in the clipboard. Columns can be ordered with headers; scroll horizontally to reach those off-screen. The bottom counters indicate visible, hidden rows, scope and filter. Close the manager doesn’t stop a Job.
Jobs menu
| Command | Use and result |
|---|---|
| Open execution details | Select a Job and choose the command, or double-click it: open the detail, with full log and execution information. |
| Rerun selected | Select a Job and open a new revision in the Runner. Check source and values; Run Job only sends a new execution. |
| Cancel running jobs | Select active Jobs and confirm the request for cancellation; a arrested Job may appear CANCELLED. |
| Copy Script ID | Copy the selected script’s registered reference to the clipboard, such as user:esempio. |
| Refresh | Reread the list and update states and counters. |
| Reconnect / recover local journal | Requires reconnection and recovery of the local Job Journal; wait for the update before consulting the state. |
| Script Library | Opens or brings the Library to the foreground. |
| Open/reconnect Job ID (local / remote) | Enter a Job ID in the window and press Open to open or reconnect details. Close closes the window without opening. For a remote ID, the calculation node is configured and enabled; if it is unknown or disabled, the details show “Execution unavailable” and the log is not available. Reconnect repeats the connection: the unavailability of the node is not equivalent to a FAILED result of the Job. |
| Close | Closes the manager; the already sent Jobs continue. |
Execution Details window
The upper panel reports identity, status, revision, hash and Job information; the lower one shows the chosen stream. Reconnect updates the connection. Cancel Job is available for an active Job and requires confirmation. When available, Open source files opens the running snapshot files in read-only language workspace: it is useful to inspect the exact code used by the Job.
Copy Script ID and Copy Job ID copy their identifiers; Copy details copies Job information. With the right button on the detail panels, the menu offers Copy selection / text and Select all. In the stream text, Select visible page selects the page displayed, Copy selection copies the selected text and Copy entire source copies the entire stored stream. Save source to file asks for a path and saves the chosen stream; here “source” indicates the selected log text source, does not change the recorded script.
The selector offers Program output, Program stderr, Diagnostic log and Legacy combined stream. Choose the stream to view it; the status bar indicates the byte range and the size. First page, Previous page, Next page and Latest page navigate pages from 64 KiB; the non-applicable buttons are disabled. Follow latest output follows updates during text production.
Publish source as Dataset asks for a new Dataset ID and publishes the selected stream as a canonical text Dataset in the Session: Press the post button only when you want to create the Dataset. The result and the ID appear in the state. This action is distinct from launching a script again.
History menu
These operations change local visibility in the list: they do not delete logs, snapshots, outputs or scientific data.
| Command | Use and result |
|---|---|
| Hide selected / Remove from visible list | Select one or more completed executions to hide them; active Jobs are excluded. |
| Restore selected | Enable Includes hidden or Hidden filter, select hidden rows and restore them in the list. |
| Hide finished in current scope | It proposes to hide the completed executions and current filter; the dialog shows the number and asks for confirmation. |
| Clear visible finished history | It offers the same hide-and-seek operation of the Jobs that are finished visible, after confirmation; it does not eliminate the results. |
View menu
Choose All for all states, Active for active ones, Finished for finished, Failed for errors or Hidden for hidden elements. The selected filter appears in the counters. Scope: All Sessions includes all Sessions; Scope: current Session uses the open one; Scope: specific Session asks for the path of the Session and Apply applies the choice. Previous page and Next page move between pages of up to 500 Jobs each. For example, with 501 results the first page shows 1 – 500 and Next page leads to the result 501; Previous page returns to the first. The command to a non-existent page is disabled. Counters reflect scope and current filters.
In the submenu Columns activates or deactivates individually reference, language, version, revision, state, started, finished, duration, session and jobid. The script name remains in the main column. The settings change the presentation, not executions.
| Option Columns | Information shown when active |
|---|---|
| reference | Registered Script ID: registered identifier of the script executed. |
| language | Language: script language. |
| version | Version: version of the script. |
| revision | Revision: revision acquired by Job. |
| state | State: current phase or outcome; see the status legend in the chapter on execution. |
| started | Started: start date and time, when available. |
| finished | Finished: end date and time, when available. |
| duration | Duration: duration of execution, when available. |
| session | Session: Session associated with Job. |
| jobid | Job ID: single execution identifier. |
Job Manager Workspace
Job Manager’s Workspace menu saves only filters, scope and layout separately from the Script Library Workspace. New starts a view with no layout selected. Save updates the current layout, and Save As / Duplicate prompts for a name for the new layout. Rename changes the name of the current layout. Open and Manager open the list of layouts. Select a row and use Open, Rename or Delete. Delete asks for confirmation and preserves Jobs, logs, outputs and hidden status. Close in the Workspace menu closes the current layout.
Settings and resource limits
Settings → Concurrency and resource policy opens the resource limits for Jobs admitted after the change. Edit the fields and press Save. The manager samples resource usage.
| Field | Meaning and unity |
|---|---|
| Concurrent jobs | Maximum number of contemporary Jobs. |
| Admission limit (active + queued) | Total limit of active Jobs and queues. |
| Threads for job | Number of threads for each Job. |
| Memory for job (MiB) | Memory for Job in mebibyte. |
| Temporary disk for job (MiB) | Temporary space for Job in mebibyte. |
| Processes for job | Number of processes for Job. |
| Time for job (seconds) | Maximum time for Job in seconds. |
In the isolated test profile, the observed defaults were 1, 64, 1, 4096 MiB, 8192 MiB, 64 and 86400 seconds, in that order. A previously configured profile may show other values. The fields require positive integers. Memory per Job is limited to 80% of physical memory. Concurrent jobs × Threads per job must be within the number of available CPUs. If a value is invalid, Save displays an error and retains the previous policy. Help → History and Workspace help summarizes Job Manager visibility and layout.