[SERVER-20438] Oplog performance on primary: efficiency of uncommitedRecordIds Created: 16/Sep/15  Updated: 02/Aug/18  Resolved: 06/Oct/15

Status: Closed
Project: Core Server
Component/s: Concurrency
Affects Version/s: None
Fix Version/s: 3.1.9

Type: Task Priority: Major - P3
Reporter: Martin Bligh Assignee: Gregory McKeon (Inactive)
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Related
related to SERVER-29274 Two Phase Drops: rename collection to... Closed
Backwards Compatibility: Fully Compatible
Sprint: QuInt A (10/12/15)
Participants:

 Description   

32 threads insert, 24 CPU box.
standalone: 1.15M insert/s
with oplog: 186K insert/s

Believed to be single threading in logOp from OplogHack mutex.
Contention from delete holding mutex whilst walking DLL is linear with list size.
Proposed fix of O(1) delete by retaining pointer handle to reference back



 Comments   
Comment by Githook User [ 25/Sep/15 ]

Author:

{u'username': u'martinbligh', u'name': u'Martin Bligh', u'email': u'mbligh@mongodb.com'}

Message: SERVER-20438: logOp refactor prior to vectorizing
Branch: master
https://github.com/mongodb/mongo/commit/0c001db908b6811c0e665e15150d1d30c47b8b5c

Comment by Githook User [ 23/Sep/15 ]

Author:

{u'username': u'martinbligh', u'name': u'Martin Bligh', u'email': u'mbligh@mongodb.com'}

Message: SERVER-20438: Stash pointer to _uncommittedRecordIds to get O(1) delete
Branch: master
https://github.com/mongodb/mongo/commit/92e38e66a79d8fc374c6c8f8faa6244f94144904

Comment by Githook User [ 21/Sep/15 ]

Author:

{u'username': u'martinbligh', u'name': u'Martin Bligh', u'email': u'mbligh@mongodb.com'}

Message: SERVER-20438: only update oplogHack once per vectored insert
Branch: master
https://github.com/mongodb/mongo/commit/c101632305f47619cd8f75429fe19621b36b727b

Generated at Thu Feb 08 03:54:13 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.