Hi,
Thanks, having set the partitionor to murmu2_random has resolved the issue with the pull queries, both where the data is and isnt in a partition that is on the same node as KSQL.
I have also (within the KSQL session) set ksql.query.pull.table.scan.enabled = true, which would allow for a full table pull query. this seems to work ok, where I have very few events to be returned but if there are more than a handfull or events it will return “Exhausted standby hosts to try.” is this related to “max.task.idle.ms” setting you mentioned on my other post?
I did have a look at what this value is meant to do and I didnt understand the explanation. At present this value is set to 0 (the default) and I am not sure how I am meant to work out what I should set this to instead, is it just a case of incrementing it by trail and error?
I am planning on marking your previous post as the solution, but I wasnt sure if that was then going to lock this post and stop me from asking about the max.task.idle.ms value
Thanks for all your help so far,
Leander