openzfs/zfs

zpool status -v doesn't report all affected files

Open

#4049 opened on Nov 26, 2015

View on GitHub
 (4 comments) (0 reactions) (0 assignees)C (9,908 stars) (1,703 forks)batch import
Type: Documentationgood first issue

Description

This is related to #4047, but warrants a separate bug. During this event, I had following situation:

  • see errors: 6331 data errors, use '-v' for a list in output of zpool status
  • run zpool status -v, get a list of files, remove them
  • run zpool status -v again to verify that all lists are replaced with hex ids (what are those exactly btw? :)
  • run zpool scrub, wait for it to finish
  • find a smaller number of data errors in zpool status -v output, containing paths to the same files as the ones referenced before, but present in a snapshot.

By "same" I mean essentially files that were created on a particular zfs volume on timestamp X, then snapshotted a number of times, without ever being modified. As I understand, in this case you'd have multiple paths pointing at the same data structure describing the contents of the file.

After chatting about this on IRC, my suspicion was confirmed - while scrub is finding all of those references to corrupted data, it most likely displays only a single path. In my case, I went through 5 scrub/cleanup cycles before I've noticed this pattern.

Contributor guide

Tech stack
c
Domain
clibackend
Issue type
bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
4
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
3-5 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
needs investigation
Prerequisites
Understanding of ZFS internalsExperience with C programmingFamiliarity with OpenZFS codebase
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
15
Research direction
Investigate the ZFS error reporting code in the zpool command to understand how it lists affected files during scrub. Look at the linked issue #4047 for related context. Examine the code in cmd/zpool/zpool main.c and module/zfs/spa.c to trace the error reporting logic. Discuss on IRC or with maintainers to clarify expected behavior.