Project

General

Profile

Actions

Bug #52599

closed

[pwl] flush requests are dispatched in advance

Added by CONGMIN YIN over 2 years ago. Updated about 2 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
Target version:
-
% Done:

0%

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

Description

If the user sends write1 write2 write3 write4 flush write5, and the depth is 2, it may become write1 write2 write3 flush write4 write5. Although this will not cause errors in normal circumstances, the semantics has changed from 1234 | 5 to 123 | 45.
https://github.com/ceph/ceph/pull/40208 want to fix https://tracker.ceph.com/issues/52277, result in the flush request be dispatched in advance.

expect: revert PR#40208 , re-fix tracker#52277.


Related issues 2 (0 open2 closed)

Related to rbd - Bug #52277: [pwl] IO hang when the single IO size * io_depth > cache sizeClosedCONGMIN YIN

Actions
Copied to rbd - Backport #53141: pacific: [pwl] flush requests are dispatched in advanceResolvedDeepika UpadhyayActions
Actions

Also available in: Atom PDF