Skip to content

Split mnt job into more granular items #51

@marcopeg

Description

@marcopeg

Right now, by running the job named mnt from the fetchq.jobs queue, the following functions are called on a single queue:

  • fetchq.mnt_mark_dead()
  • fetchq.mnt_reschedule_orphans()
  • fetchq.mnt_make_pending()

By running so much stuff all together, we increase the possibility for introducing a dead lock on the table as we are updating multiple rows at the time, and also dealing with stats updates.

We can introduce specific lines in the jobs table to run each of them in isolation, reducing the single maintenance time hit, reducing the probability for dead locks

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions