[SERVER-4392] split network thread out from worker thread for secondaries Created: 30/Nov/11  Updated: 11/Jul/16  Resolved: 18/May/12

Status: Closed
Project: Core Server
Component/s: Replication
Affects Version/s: None
Fix Version/s: 2.1.2

Type: Improvement Priority: Major - P3
Reporter: Eliot Horowitz (Inactive) Assignee: Kristina Chodorow (Inactive)
Resolution: Done Votes: 2
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Depends
is depended on by SERVER-5685 Have multiple threads applying oplog ops Closed
Participants:

 Description   

Exhaust is probably the right way to do it.



 Comments   
Comment by auto [ 16/May/12 ]

Author:

{u'login': u'kchodorow', u'name': u'Kristina', u'email': u'kristina@10gen.com'}

Message: Added comments and some better naming SERVER-4392
Branch: master
https://github.com/mongodb/mongo/commit/cf6a249630605d3e9f1bcb7ab83172413970a890

Comment by auto [ 16/May/12 ]

Author:

{u'login': u'kchodorow', u'name': u'Kristina', u'email': u'kristina@10gen.com'}

Message: Stop notifier thread from spinning SERVER-4392
Branch: master
https://github.com/mongodb/mongo/commit/53671c4877320460a64949b214f0585dd8b5c7ff

Comment by auto [ 16/May/12 ]

Author:

{u'login': u'kchodorow', u'name': u'Kristina', u'email': u'kristina@10gen.com'}

Message: Made repl use BlockingQueue SERVER-4392
Branch: master
https://github.com/mongodb/mongo/commit/235ea9fd797d8c41422f65602b307a39ded59cdb

Comment by auto [ 16/May/12 ]

Author:

{u'login': u'kchodorow', u'name': u'Kristina', u'email': u'kristina@10gen.com'}

Message: cleanup includes and tests SERVER-4392
Branch: master
https://github.com/mongodb/mongo/commit/39629ff51d5f2b2361328681e63b742e601f307a

Comment by auto [ 16/May/12 ]

Author:

{u'login': u'kchodorow', u'name': u'Kristina', u'email': u'kristina@10gen.com'}

Message: Separate thread for fetching oplog ops SERVER-4392
Branch: master
https://github.com/mongodb/mongo/commit/5ba5abbb01583e637e7d380ab869bf4d695848af

Comment by auto [ 16/May/12 ]

Author:

{u'login': u'kchodorow', u'name': u'Kristina', u'email': u'kristina@10gen.com'}

Message: Better naming SERVER-4392
Branch: master
https://github.com/mongodb/mongo/commit/adfe6f647fcbf7737c81fa747d8bf3599cc7d84c

Comment by auto [ 16/May/12 ]

Author:

{u'login': u'kchodorow', u'name': u'Kristina', u'email': u'kristina@10gen.com'}

Message: Move w tracking to its own thread SERVER-4392
Branch: master
https://github.com/mongodb/mongo/commit/4d2b8d30f3f730a42cfaba7697b4acf19e339612

Comment by auto [ 16/May/12 ]

Author:

{u'login': u'kchodorow', u'name': u'Kristina', u'email': u'kristina@10gen.com'}

Message: Add rollback check to oplog reading file SERVER-4392
Branch: master
https://github.com/mongodb/mongo/commit/e6e92ac7aaf625c0fa12455118e522ae77ce2b51

Comment by auto [ 16/May/12 ]

Author:

{u'login': u'kchodorow', u'name': u'Kristina', u'email': u'kristina@10gen.com'}

Message: Move oplog-related helpers to their own file SERVER-4392
Branch: master
https://github.com/mongodb/mongo/commit/83447912f127019bf4e0270cdd3bec8fb9e483ea

Comment by auto [ 24/Apr/12 ]

Author:

{u'login': u'dwight', u'name': u'Dwight', u'email': u'dwight@10gen.com'}

Message: SERVER-4392 remapprivateview() more often
groupCommitWithLimitedLocks cannot do a remap as it lets people write for part of its work. every Nth time
because of this we do a "non-limited" commit so that we can remap.

on that invocation, if there was no writes pending, we would return early and not call remap. thus in theory
if you had writes on the other times but not on the N times, remap wouldn't get called enough.

in reality, this might be impossible statistically to happen. however, this variation feels safer. please
review. note there is some inefficiency here : we upgrade() our lock and then in remap we may decide to do
no work. however, that will mean the performance is predictable : if you suddenly do trivial # of writes,
it doesn't start behaving wildly differently here. that would be the argument for keeping it this way.

there is a separate issue with insufficient remapping that has to do with commitIfNeeded. that is still
outstanding.
Branch: master
https://github.com/mongodb/mongo/commit/a486732a251c80b4df9a04be3ef0ef34e7167e3b

Comment by auto [ 14/Mar/12 ]

Author:

{u'login': u'dwight', u'name': u'dwight', u'email': u'dwight@10gen.com'}

Message: SERVER-4392 fix bug introduced in test
Branch: master
https://github.com/mongodb/mongo/commit/00f39a03ac068386464e96a5f44da7adb6d8062a

Comment by Dwight Merriman [ 19/Dec/11 ]

right, QueryOption_Exhaust is the way to go. we use that in clone collection already

Generated at Thu Feb 08 03:05:51 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.