Jordan Miller

That worked and makes total sense. Thanks for the tip. In case you are curious the real production statement isroughly ~600 lines long and creates multiple temp tables. The output looks like this: (4955 rows affected) (4885 rows affected) (35 rows affected) (30 rows affected) (5 rows affected) (164510 rows affected) (6214 rows affected) (40296 rows affected) (25326 rows affected) (78767 rows affected) (2040 rows affected) (2009 rows affected) (667 rows affected) (13059 rows affected) (4947 rows affected)...

That worked and makes total sense. Thanks for the tip. In case you are curious the real production statement isroughly ~600 lines long and creates multiple temp tables. The output looks like this: (4955 rows affected) (4885 rows affected) (35 rows affected) (30 rows affected) (5 rows affected) (164510 rows affected) (6214 rows affected) (40296 rows affected) (25326 rows affected) (78767 rows affected) (2040 rows affected) (2009 rows affected) (667 rows affected) (13059 rows affected) (4947 rows affected)...

Thank you both for the replies. Sadly the data source we are reading from is read only for schema changes, so deploying things like new SPs and View's, while a great idea is not really an option. Plus for many things a quick and simple temp table inline is just a REALLY useful tool. As for multiple select statements, if I remove the second one it will still blow up when the "INTO #peopletemp" statement is used. Here is the stack trace java.lang.NullPointerException: rs must not be null at java.base/java.util.Objects.requireNonNull(Unknown...

I'm trying to define Source SQL executing against a MS SQL Server that selects INTO a temp table and then selects from that table. The following is a simplified version of what I'm trying to do that recreates the issue: SELECT [name] ,[birthday] INTO #peopletemp FROM [server].[dbo].[people] SELECT * FROM #peopletemp If I remove the INTO statement and the second SELECT the basic query runs fine. If it matters I'm adding a new report through the web UI.

That's helpful thank you, I'll have to track that setting down. I assume we would need to use a shared volume between the instances for things like template files correct?

We are looking to do something VERY similar. We quickly got a docker image working and now we would like to be able to auto scale the containers up and down based on load (all sitting behind a single load balancer). The basic requirement for this is that each instance needs to not cache and essentially be stateless (aside from any job that it is currently running on a particular node). @ecerichter , did you find that the setting recommended by @tanyona worked as expected and gave you a solid HA ...