We are experiencing problems when trying to copy a workflow with many associated common components to knime server. This is either when we use a manual copy or an API copy. Other knime clients (other machines) hang and cannot connect during the copy operation.
Any suggestions on how to improve performance during copy operations for a large workflow with many common components?
How “large” are we talking in your case?
When you upload a workflow to KNIME Server it will be extracted into a temporary folder and then moved into the server repository. This is usually a fast operation if both the temporary directory and the server repository are on the same partition. During this process access to other clients is locked so that they don’t see an inconsistent state.
Therefore you either have a really large workflow (tens of thousands of files) or your file system performance is quite low.
well, its not that big. but have a lot of common components. One thing we are seeing is that the snapshots are being copied as part of the “copy” operation. wonder if we can opt out of snapshot copy as part of the “copy” operation
Aha, so it’s not an upload (“copy to KNIME Server”) but a server-side copy. This is indeed more expensive because it’s a real copy and not just a move.
There is no possibility to exclude the snapshots when performing a server-side copy.
good to know. thanks
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.