Skip to content

Eventually consistent batch scanner retries too aggressively on failure #6313

@keith-turner

Description

@keith-turner

Describe the bug

When an eventual batch scanner encounters a RPC error it will retry immediately in some cases (maybe all, not sure) . This aggressive retry could exacerbate a problem in the distributed system. The cause of the problem is that the batch scanner code asks the client side scan server plugin for retry time and that plugin always returns zero. Immediate batch scanners do not have this problem, they do some custom backoff.

Versions (OS, Maven, Java, and others, as appropriate):

  • Affected version(s) of this project: [e.g. 1.10.0]
  • OS: [e.g. CentOS 7.5]
  • Others:

To Reproduce

Noticed this problem when running new test added in #6307 and observing client side test logs. Saw the batch scanner retrying a lot before it timed out.

Expected behavior

Eventual batch scanners do some sort back of when there are failures. Should also look into the scanner behavior for this case.

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugThis issue has been verified to be a bug.

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions