-
Notifications
You must be signed in to change notification settings - Fork 8.4k
Possible deadlock in StorageKafka #68360
Copy link
Copy link
Closed
Labels
comp-message-queuesMessage queue integrations (Kafka, RabbitMQ, NATS table engines for stream ingestion/egress).Message queue integrations (Kafka, RabbitMQ, NATS table engines for stream ingestion/egress).potential bugTo be reviewed by developers and confirmed/rejected.To be reviewed by developers and confirmed/rejected.
Metadata
Metadata
Assignees
Labels
comp-message-queuesMessage queue integrations (Kafka, RabbitMQ, NATS table engines for stream ingestion/egress).Message queue integrations (Kafka, RabbitMQ, NATS table engines for stream ingestion/egress).potential bugTo be reviewed by developers and confirmed/rejected.To be reviewed by developers and confirmed/rejected.
Type
Fields
Give feedbackNo fields configured for issues without a type.
There is a very strange failure of kafka integration tests. Basically at some point (around 15:05:16) the logs just stopped. I suspected that something happened with with kafka, but the kafka node doesn't have any logs. In clickhouse I only see the queries from teardown, but even the
DROP DATABASEqueries time out. In the mix ofCREATE TABLEandDROP DATABASEqueries the last created table is the old kafka engine, so I think the old kafka engine deadlocked in some way. And it makes sense, comparing the last two (which used the old kafka engine)DROP DATABASEquery, it is clear that we have a deadlock inStorageKafka. I am suspecting either #66311 PR or my Kafka ZooKeeper PR #57625 to be the cause.