Similarity search error: "Queue size must be as least as large as running instance count"

KNIME Analytics Platform
1

Hi All, I have seen a similar post form 2015 with no answers here: Struggling with the Similarity Search node - maybe misplaced?

The error persists however: I read in a csv, define a distance measure based on multiple columns, no problems with configuring the nodes, and then… “Execute failed: Queue size must be as least as large as running instance count: 100 vs. 144”

The same workflow did not produce this error on a different VM with a different amount of cores (I’m using AWS instances), but I am currently trying to reproduce that and try to find a configuration on which it works…

Any ideas?

2

Hi

sorry you couldn’t find an answer at the previous forum post.
Could you give as a short example workflow (ideally with data) where this happens?

Best
Alice

1 Like
3

Hi pogacsa

just checking, could you reproduce the error?

Best
Alice

4

Hi Alice, I could only produce the error on an AWS machine with 96 cores. As soon as I ran the same workflow with only 32 cores, all was fine. It is only produced by the similarity search node. Have you tried running this node on very large machines?

5

Hi pogacsa

we followed up on this, and indeed there was an issue once you’d use a machine with >64 cores. It should be fixed now. Thanks a lot for reaching out! :slight_smile:
Please let us know if the problem should continue to occur.

Best
Alice

2 Likes
6

Thank you! :slight_smile: I will test asap.

7

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.