"Results have expired, rerun the query if needed"

[jerryleooo]

Describe the bug:

Global variables are not shared between different gunicorn workers.

hue/desktop/libs/notebook/src/notebook/connectors/sql_alchemy.py

Line 90 in b5d126e

CONNECTIONS = {}

Steps to reproduce it?

1. Use the docker-compose to up the service, with following configurations:

gunicorn_work_class=eventlet
gunicorn_number_of_workers=8

2. configure an interpreter with Athena/Trino/Presto + sqlalchemy
3. use 2 different browser tabs to submit queries in editor, and you may see the Results have expired, rerun the query if needed. error

Hue version or source?

open source 4.10.0

[sbbagal13]

I too observed similar issue with Trino

[bjornalm]

Hi @jerryleooo, thanks for reporting this. Do you know if this happens with other interpreters as well?

@Harshg999 can you have a look?

[jerryleooo]

Hi @bjornalm I haven't tried other interpreters, later i may try jdbc and update here

[github-actions]

This issue is stale because it has been open 30 days with no activity and is not labeled "Prevent stale". Remove "stale" label or comment or this will be closed in 10 days.

[sbbagal13]

Is it resolved in recent release or do we need to change the interpreter for this ?
CC: @bjornalm @Harshg999

[vpattar]

I too observed similar issue with Trino, which is connected through a VPC endpoint.

Is there a known workaround to this?

[jerryleooo]

@vpattar now I use gunicorn_work_class=sync and gunicorn_number_of_workers=1, and running 3 instances behind a LB

[Harshg999]

With Gunicorn now, we need to relook and scope how to share global variable state. I've created an internal roadmap item for this issue and will take it up for subsequent releases.
Thanks for flagging this @jerryleooo.