...
Since there can be several steps to retrieving an output file, this information is made available by clicking on the Information button in the toolbar.
Local output files
Output files may be read in one of three ways, attempted in the following order:
- from a log server (if defined)
- directly from disk (optional)
- from an ecFlow server
ecFlowUI will always attempt to obtain a file from the log server if it is defined for this task. Then, optionally, it attempts to read the file directly from disk if it is accessible. Finally, if the file has not yet been read, ecFlowUI will request the file from the ecFlow server.
This behaviour tries to avoid the situation where the wrong file is read because the same path exists on the machine where the task is running and on the machine where ecFlowUI is running, but they are not the same file. As an example, consider that the machine that ran the job wrote its output into a file called /tmp/myjob.1, where /tmp is a file system local to that machine. If a different file of the same name exists on the machine where ecFlowUI is running, this could be read instead of the 'real' output file.
If you know that this could be the case, you should go to the Server Settings and deactivate the option Read files from disk when appropriate.
| Anchor | ||||
|---|---|---|---|---|
|
...