Project

General

Profile

Feature #820

bufferlist memory pool

Added by Sage Weil over 9 years ago. Updated over 9 years ago.

Status:
Closed
Priority:
Normal
Assignee:
Category:
-
Target version:
% Done:

0%

Source:
Tags:
Backport:
Reviewed:
Affected Versions:
Pull request ID:

Description

Investigate putting buffers in a separate memory pool.

History

#1 Updated by Sage Weil over 9 years ago

  • Target version changed from v0.25 to v0.26

#2 Updated by Greg Farnum over 9 years ago

To elaborate: it appears when I compare the heap memory use reported by massif* and the resident memory usage reported by top that we are continuing to have serious issues with memory fragmentation. tcmalloc improved this situation over ptmalloc but it's still not good enough for people running in RAM-limited situations and our heap usage is low enough that it should allow proper use.

[*] or tcmalloc heap profiler.

#3 Updated by Greg Farnum over 9 years ago

  • Assignee set to Greg Farnum

#4 Updated by Greg Farnum over 9 years ago

  • Status changed from New to In Progress

I did a prototype of this, and in brief testing it reduced resident memory usage a little bit, but not by enough. More experimentation might improve results; but after looking at the test data some and discussing with Sage, #834 definitely comes first!

#5 Updated by Greg Farnum over 9 years ago

  • Status changed from In Progress to Closed

So, tried this out and it looked like maybe it helped a little bit, but it wasn't clear. After running more tests with higher tcmalloc output, and examining the proc/maps, it's starting to look like memory use issues aren't actually with the heap but...something else. I don't think this is worth much more looking into.

#6 Updated by Sage Weil over 9 years ago

  • translation missing: en.field_story_points set to 2
  • translation missing: en.field_position set to 1
  • translation missing: en.field_position changed from 1 to 538

Also available in: Atom PDF