I have a question. “in progress” queries sometimes remain after clicking Cancel button in admin page.
It is because “query_task_trackers:in_progress” in redis remain even if Celery tasks we canceled . I don’t know how to reproduce these data, but should I flushall redis data?

1 Like

I’m having this exact problem, with two queries stuck in executing query. I cannot cancel them and I cannot run them again because the system evidently things they are running. How can I flush or cancel or otherwise remove them from the queue?

It’s worth noting that this seems to be a similar problem:

and this:


After multiple tries of redis-cli flushall and rebooting the server, the queries finally cleared.

I am repeatedly having the same issue. I wish the cancel button actually worked since the workaround (to flush redis and reboot server through the terminal) requires that either every data analyst has to ask for this to the systems engineer or every analyst is also a systems engineer.
Is there another way?

The current Queries Queue tab has issues where “phantom” tasks might get stuck there and you won’t be able to ever remove or cancel them.

For this reason the following change was introduced:

It should be more reliable as it reads Celery’s state directly.