Project

General

Profile

Bug #38176

Unable to recover from ENOSPC in BlueFS, WAL

Added by Adam Kupczyk about 5 years ago. Updated about 5 years ago.

Status:
Won't Fix
Priority:
Normal
Assignee:
Target version:
-
% Done:

0%

Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

It is possible to insert so much OMAP data into objects that it will overflow storage and cause ENOSPC when rocksdb tries to flush WAL, and needs to extend bluefs space.
This is fatal error as there is really less space then required.

This bug happened as result of testing of bug #36268.
How to replicate this bug:
https://drive.google.com/file/d/1KJAmuz2YGoL17UYEmd6VUegBMABk6gmG

History

#1 Updated by Kefu Chai about 5 years ago

  • Status changed from New to Fix Under Review
  • Assignee set to Adam Kupczyk
  • Pull request ID set to 26276

#3 Updated by Neha Ojha about 5 years ago

  • Status changed from Fix Under Review to Won't Fix

We decided to not fix this.

#4 Updated by Adam Kupczyk about 5 years ago

Solution proposal: https://github.com/ceph/ceph/pull/26276

It was decided not to implement it at this stage.

Also available in: Atom PDF