[SERVER-5142] double check after lock that ShardChunkManager in trySetVersion needs to be rebuilt Created: 29/Feb/12  Updated: 11/Jul/16  Resolved: 06/Mar/12

Status: Closed
Project: Core Server
Component/s: Sharding
Affects Version/s: None
Fix Version/s: 2.0.4, 2.1.1

Type: Bug Priority: Critical - P2
Reporter: Greg Studer Assignee: Greg Studer
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Depends
Related
related to SERVER-5199 more granular locking for config serv... Closed
Operating System: ALL
Participants:

 Description   

... otherwise creating a new ShardChunkManager is wasteful and can spam the config server if many threads are blocked on the mutex.



 Comments   
Comment by auto [ 06/Mar/12 ]

Author:

{u'login': u'gregstuder', u'name': u'Greg Studer', u'email': u'greg@10gen.com'}

Message: SERVER-5142 use ticket holder instead of mutex to allow 3 parallel refreshes, with 30s timeout
Branch: v2.0
https://github.com/mongodb/mongo/commit/8983591f83c0817d0bf46a29d338db910399d8e7

Comment by auto [ 06/Mar/12 ]

Author:

{u'login': u'gregstuder', u'name': u'Greg Studer', u'email': u'greg@10gen.com'}

Message: SERVER-5142 double-check in lock for new version before recreating ShardChunkManager
Branch: v2.0
https://github.com/mongodb/mongo/commit/edc6a4892119009b514202e4d0c97d5c3bf54e1d

Comment by auto [ 06/Mar/12 ]

Author:

{u'login': u'gregstuder', u'email': u'greg@10gen.com', u'name': u'Greg Studer'}

Message: SERVER-5142 use ticket holder instead of mutex to allow 3 parallel refreshes, with 30s timeout
Branch: master
https://github.com/mongodb/mongo/commit/c3443fc81194882c4b2673d0b8672b5358ba48e6

Comment by auto [ 04/Mar/12 ]

Author:

{u'login': u'gregstuder', u'name': u'Greg Studer', u'email': u'greg@10gen.com'}

Message: SERVER-5142 double-check in lock for new version before recreating ShardChunkManager
Branch: master
https://github.com/mongodb/mongo/commit/231ffceb86cb2f5b2c355f3e43e6eb524c244bec

Comment by Greg Studer [ 02/Mar/12 ]

http://codereview.10gen.com/7490001/

Comment by Greg Studer [ 01/Mar/12 ]

Not this time . Related to locking around objects that hit the config server - we need to make this more efficient.

Comment by Andy Schwerin [ 01/Mar/12 ]

I could be wrong, but every time I see "double-check locking," I assume it should be related to the "startup initialization solution" issue

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