Skip to content

ILM uses end-time of SLM snapshot when waiting for snapshot #73357

@DaveCTurner

Description

@DaveCTurner

The ILM wait-for-snapshot step aims to ensure that we have captured a snapshot of the final state of the index before proceeding. It does this by waiting for the last successful SLM snapshot to have a timestamp after the corresponding ILM phase transition time. However the SLM snapshot timestamp records when the snapshot completed, which will be some time after the index state was captured, so we cannot be sure that it had reached its final state before being snapshotted.

I believe instead we should wait for an SLM snapshot which started after the ILM phase transition, rather than one which completed after this transition.

Relates #73356 regarding the use of wall-clock timestamps in this area.
Relates #57809 regarding other guarantees offered by the wait-for-snapshot step.

Metadata

Metadata

Assignees

No one assigned

    Labels

    :Data Management/ILM+SLMDO NOT USE. Use ":StorageEngine/ILM" or ":Distributed Coordination/SLM" instead.>bugTeam:Data Management (obsolete)DO NOT USE. This team no longer exists.

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions