NDT rebuild timing out at 5 mins after sql change to underlying table

(Josh Hatfield) #1

I am changing the way a table is joined to the table one of my ndts are based on. The issue is when trying to run anything in the NDT after I make the change, Looker seems to recognize the change and force a total rebuild.

The sql for the rebuild is very long and when trying to run it I get ‘error running query’ at 5 minutes each time. Is there a way to extend this 5 minutes? I am using redshift through workbench/j.


(Izzy) #2

Hey Josh!

5 minutes does seem like a pretty short timeout. Looker actually doesn’t enforce any timeouts, technically (though our default hosted server will time out queries at 60 minutes if they’re run in the browser).

That’s to say, this sounds like there might be a timeout setting on your redshift DB, or on some piece of your stack between looker and your DB (a load balancer or proxy). I’d check the db for a timeout setting first, and then work down the chain to Looker. Keep us posted on what you find!


(Josh Hatfield) #3

was a load balancer issue. eventually figured that out, was able to change in ec2 management console.

Thanks for your response!