Skip to content

Request hang when disk is corrupted #486

@zuston

Description

@zuston

When Riffle detects disk corruption, it marks the server as unhealthy. However, the await-tree log still shows hung requests reading from the broken disk. In fact, we already have an I/O timeout layer: if a request’s duration exceeds the configured threshold, it should be stopped.

Image

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions