Open jakcodex opened 10 months ago
These columns were added recently. It is possible that one of the broker nodes is running an older version despite the upgrade. Do confirm that the broker throwing these errors is indeed running on 27.
This issue has been marked as stale due to 280 days of inactivity. It will be closed in 4 weeks if no further activity occurs. If this issue is still relevant, please simply write any comment. Even if closed, you can still revive the issue at any time or discuss it on the dev@druid.apache.org list. Thank you for your contributions.
Affected Version
27.0.0
Description
Apache Druid is deployed in a highly available cluster in AWS using Amazon Linux 2023 x86_64 AMI consisting of three-node master-with-zk cluster and an assortment of data/query nodes using MySQL-based metadata and S3 deep storage. The cluster presently has no segments or configured tasks.
Management UI running on a 27.0.0 server experiences intermittent failures for requests to
/druid/v2/sql
.The Services page makes a request with the following payload:
Which sometimes get the error response:
400 Bad Request
Unified Console home page makes the following request:
Which sometimes get the error response:
400 Bad Request
Strangely, when watching the automatic refresh requests on the Services page, the error occurs exactly every other request consistently.
Errors are being logged in the corresponding broker's log file.
With all cluster nodes installed with 26.0.0 this issue is not present. With all cluster nodes installed to 27.0.0 the issue is present. If you then downgrade a query/router node to 26.0.0 the issue is no longer present on that specific node.