This message was deleted.

interesting. do you see any errors in historical or broker logs?

I don't think there were, but I since removed the pods. I'm right now trying again from scratch and so far it looks like this time it works....

certain metadata fields (including

num_rows

) are populated in the background and take some time to appear when a segment is first loaded or when a Broker restarts. It doesn't affect regular queries, only metadata queries. You might have run into this

@Gian Merlino I think this also had to do with shared volumes. https://apachedruidworkspace.slack.com/archives/C0309C9L90D/p1662491715810719?thread_ts=1662491715.810719&cid=C0309C9L90D

ah, i see