are there any example workflows for using the Table Based Call Remote Workflow nodes with the server? I have a workflow on the server I want to call, but when I try to configure the node, I get a message to say that the node āCould not locate any of the required nodes (read node description) on the server. Please check that the necessary nodes are installedā
Thanks, Ivan. The Call Remote Workflow node wonāt let me apply any advanced settings - the dropdowns are greyed out as a result of the error message. Iāve tried to test the process again using a simpler workflow on the server, but I get the same problem. Executes perfectly when saved locally, but says it canāt find the nodes when on the server. As far as I can see the server connection node is set up correctly, so Iām stumped.
The workflow example linked at the nodepit site doesnāt have these nodes in it, unfortunately, so itās not any help in this case.
I see. I just tried it and works for me only I donāt have Server
It is said the workflow path needs to be absolute. Is it in your case? As long as you have proper nodes in your calling workflow things should work correctly.
Yeah, it works if I have the both workflows saved locally - itās something to do with the server, I think. The container nodes are definitely in the workflow. The dialog box pre-populates a list of available workflows on the server, so the path should be correct. The connection to the server is working according to the test function in the connection node. Itās weird.
According to the help dialog, itās the absolute path - if itās not, thatās a bug in the node setup. The node can find the workflows on the server, itās just not recognizing the container nodes within them. And I only have one server
This is a new node, and it looks like we still need to do a bit of work on the documentation. I know that weāre planning to add a couple of other improvements to the node for the next release.
To use the node on the KNIME Server you actually have two options. You can explicitly setup a KNIME Server connection as you have done. That would be handy if you want to run a bunch of calculations locally, and then leverage server compute for part of the calculation. However, if you want to use the workflow on server, or analytics platform you can actually use a relative path like ā¦/path/to/workflow that would be the same for the workflow locally as on the server.
It sounds like there might be something funny going on. Are you able to share the simple workflow with me. Also helpful is the KNIME Server version and the KNIME AP version.
Iāve just been reproducing that behaviour on my KNIME Server. The issue stems from the version of the server that you are using. KNIME Server 4.6.x uses the 3.5 executor. AP 3.5 and the corresponding executor do not contain the relevant nodes (Container Input (Table) and Container Output (Table)). Thatās because we only released those nodes with AP 3.6. That means that youād need to update to KNIME Server 4.7.x to test the functionality. Iāve opened a ticket to see if we can improve the error message.
The AWS version will be published in the next few weeks. Weāre waiting for the AP 3.6.1 release, and will also add in a couple of performance fixes for the KNIME Server that weāre currently verifying.