[SERVER-17600] Memory leak during stress - insert only workload to cap collection Created: 16/Mar/15  Updated: 26/Sep/15  Resolved: 26/Sep/15

Status: Closed
Project: Core Server
Component/s: Storage, WiredTiger
Affects Version/s: None
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Eitan Klein Assignee: Unassigned
Resolution: Incomplete Votes: 0
Labels: 28qa
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Attachments: PNG File Process_Private_Bytes0.png     PNG File Process_Private_Bytes0.png     Text File ServerStatus.txt     Text File diff4.txt    
Issue Links:
Depends
Operating System: Windows
Participants:

 Description   

db version v3.0.1-pre-
git version: 0cbcdfd5b4cd4961b448f1a35fd0d4ea19629b93

Environment:

  • Single machine
  • Windows 2008
  • Each instance uses 1GB RAM of cache

Workload:

  • Used hammer.mongo to do insert only workload to cap collection of 1Gbyte
  • Attached UMDH output that highlight the allocations that remained in memory


 Comments   
Comment by Michael Cahill (Inactive) [ 09/Jun/15 ]

This may be fixed by the switch of memory allocator on Windows to tcmalloc.

Comment by Martin Bligh [ 09/Jun/15 ]

Will see if it reproduces on Linux

Generated at Thu Feb 08 03:45:01 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.