Good afternoon everyone, it’s a pleasure to greet you.
I would like to kindly request a potential improvement from the developers regarding KNIME’s behavior in certain unexpected scenarios.
Currently, if our PC or server experiences a connectivity issue or, for some other reason, KNIME closes unexpectedly while a workflow is running or becomes unresponsive, the workflow can end up being internally damaged. Ideally, in such cases, the system should avoid applying any changes and instead preserve the last saved version of the workflow, to prevent corruption.
I understand that keeping regular backups is the best practice. However, in reality, it’s common to make updates to a workflow without saving immediately, and in those cases, there’s a high risk of losing work.
Thank you very much for your understanding and for the great work you do.
Best regards.
The workflow should only be saved when you initiate a save action explicitly (Ctrl/Cmd+S, close and “Save” in dialog, we even try to detect an unresponsive UI and offer to save your work manually).
Any corruption occurring is the consequence of a bug, and bugs are something we continuously try hard to avoid, but unfortunately cannot be made impossible when enhancing the software with so many new features and capabilities.
Do you have a specific instance in which your workflow on disk got corrupted? We could follow up on that and investigate what went wrong to avoid it in the future.
In general, backups are a good idea not only to guard against bugs in software, but also in case a change was saved that should not have been. On Community Hub, we now offer to save workflow versions, in case that would be an option for you and your team. Analytics Platform 5.5, which we just released, makes it really easy to go back in time right from the versioning panel in AP, where you can switch between workflow versions and compare what changed.
My workflow is quite large and consists of numerous nodes, which I organize into components and metanodes to make it more manageable. However, I’ve noticed a problematic behavior: when the system freezes, the processes within these components or metanodes tend to become corrupted. As a result, when I reopen the workflow, several nodes are missing and only fragments of the original design remain.
What’s most concerning is that in such cases, the workflow is automatically saved without asking whether I want to keep the latest changes. This leads to the loss of previous work, with no way to recover it.
My suggestion is that the tool should avoid performing automatic saves in the event of unexpected shutdowns or internal errors—or at least provide an option to disable this feature. This would help prevent data loss in complex workflows.
This does sound rather unexpected (both the metanodes and component missing contents and the corrupted state being saved automatically). Analytics Platform does not include the function to automatically save workflows in case of freeze or another problem. It tries to detect an unresponsive user interface and asks the user if they want to save or not (but never does it automatically).
Though I have not personally used that ever. Do you by chance have that activated? This could give us a clue what exactly is causing your workflows to lose content.
Which version of Analytics Platform are you using?
Help us to help you.
Take a look at how to report your findings
Prevencion is a way to go. I recomend making version backups of your wfs.
…it’s common to make updates to a workflow without saving immediately, and in those cases, there’s a high risk of losing work.
The must:
-Perform complete backups of all critical data and custom code.
-Verify backup integrity in advance.
-Prepare a rollback plan in case the update fails.
-Ensure procedures are in place to quickly restore previous versions.
-Set up redundancy (e.g., secondary emvironment ) to minimize downtime.
-Test the updates in a non-production (staging) environment, NEVER in production.
In the KNIME Classic User Interface, the option is disabled. However, under the General section, the Workbench Save Interval (in minutes) option is enabled and set to 5 minutes.
I am currently using version 5.4.4. However, a colleague who is using version 5.5 has also experienced the same issue.
