26–28 Mar 2014
CERN
Europe/Zurich timezone

AFS Salvager and Large Volumes

28 Mar 2014, 10:00
30m
31/3-004 - IT Amphitheatre (CERN)

31/3-004 - IT Amphitheatre

CERN

105
Show room on map

Speaker

Mr Mark Vitale (Sine Nomine Associates)

Description

The AFS salvager is one of those programs (like the restore part of backup/restore) that are only needed when something has already gone wrong, and must be set right as quickly as possible. In a perfect world, there would be no need for recovery tools like the AFS salvager; servers would always come down cleanly, and filesystems would always be in a consistent and correct state. But we don’t live in a perfect world: hardware fails, servers crash, and AFS has bugs. Undiscovered bugs are more likely to lurk in code paths that are 1) complex and 2) rarely exercised, e.g. salvaging very large AFS volumes. This presentation will discuss the AFS namei partition format and how salvager identifies and repairs damage in it. We will conclude with a real-world case study that was used as a starting point for a test suite to identify and fix remaining bugs in AFS large volume support. We'll also briefly discuss possible alternatives to the salvager.

Primary author

Mr Mark Vitale (Sine Nomine Associates)

Presentation materials