Failing to communicate with Presto after "Failed getting schema"

[gppeixoto]

Issue Summary

I believe this is likely to be a front-end issue, but its root is caused by an external Presto issue. Redash is not able to issue a new query on /queries/new after a "Failed getting schema" error, even though the connection is OK and queries run normally on the backend.

• Search for existing issues first

Steps to Reproduce

1. Connect a presto datasource to redash;
2. Confirm test connection is OK (it should issue a SHOW TABLES query to Presto)
3. Enter /queries/new, it will issue the following query to presto:

SELECT table_schema, table_name, column_name
FROM information_schema.columns
WHERE table_schema NOT IN ('pg_catalog', 'information_schema')

4. If presto query return INTERNAL ERROR and you try any query (e.g.: select * from some_table limit 1) , front-end shows a box "Query in queue… 00:00:00" that runs forever and the query is never issued by the server

The internal error that I get on Presto is tracked on: prestodb/presto#6972, and it is still an open issue. However, any query that I run directly on any presto table runs normally, including any "describe some_table".

Technical details:

• Docker image redash/redash:2.0.0.b2905 running on Kubernetes v1.5.4
• OS X Yosemite 10.10.5
• Chrome 58.0.3029.110 (64-bit)
• Presto 0.157.1

[arikfr]

It happens that fetching the schema fails for various reasons, I never saw it impact the ability to run queries. Are you sure you had the Celery workers running at the time? If you open the query window the second time, does it run the query?

[dulh]

I have the same issue above, the query get error when execution, maybe the command invalid !
I'm using Presto v0.197