TimeoutException on all requests (async deadlock ?)

Description

When deploying my app (which runs on cassandra) in production, I started to notice TimeoutExceptions throwing from time to time, while none of my 3 cassandra nodes had any load.

It started to happen more and more often as my server load was increasing.
My server now rejects almost every request because of these timeouts... Which is, as you can imagine, a bit annoying

I managed to reproduce it with a simple example, which (on my local instance of cassandra or on my production cluster) results systematically in requests throwing TimeoutExceptions:

(code also attached in zip)

After a bit of investigation, it seems that there is an async deadlock somewhere.

I'm not clear about why it happens, though.

Am I missing something ?

Thanks !

Environment

None

Attachments

2

Activity

Show:
Not a Problem

Details

Assignee

Reporter

Labels

Reproduced in

Components

Affects versions

Priority

Created March 22, 2017 at 5:34 PM
Updated January 16, 2022 at 8:04 PM
Resolved March 23, 2017 at 10:35 AM