Project

General

Profile

Bug #13127

ObjectCacher: coalesce multiple writes to a single object into one IO

Added by Greg Farnum almost 4 years ago. Updated over 3 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
-
Category:
ObjectCacher
Target version:
-
Start date:
09/16/2015
Due date:
% Done:

0%

Source:
Development
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:

Description

Right now, if the ObjectCacher has dirty data at multiple offsets of a single object, each contiguous region becomes its own write Op. Some usage patterns make this much slower than if it built up an Op containing all the writes for the object in question. For instance, the ffsb and fsync testers we run against ceph-fuse (and krbd?) can generate >100,000 OSD requests on a 64-object file. (This is #13126 for ceph-fuse.)

History

#1 Updated by Zheng Yan almost 4 years ago

  • Status changed from New to In Progress

#2 Updated by Zheng Yan almost 4 years ago

  • Status changed from In Progress to Need Review

#3 Updated by Greg Farnum over 3 years ago

  • Status changed from Need Review to Resolved

Also available in: Atom PDF