Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | SERVER-23971 Clang-Format code | Mark Benvenuto | 2016-05-28 | 1 | -1/+1 |
| | |||||
* | SERVER-23271 Add keysInserted and keysDeleted metrics for CRUD ops | James Wahlin | 2016-04-21 | 1 | -1/+2 |
| | |||||
* | Revert "SERVER-23271 Add keysInserted and keysDeleted metrics for CRUD ops" | Benety Goh | 2016-04-21 | 1 | -2/+1 |
| | | | | This reverts commit 6bbaee174447ee1c9177c72bdd07f050ab07e901. | ||||
* | SERVER-23271 Add keysInserted and keysDeleted metrics for CRUD ops | James Wahlin | 2016-04-21 | 1 | -1/+2 |
| | |||||
* | SERVER-22178 Move calls that cannot throw out of try/catch block. | Charlie Swanson | 2016-03-25 | 1 | -1/+1 |
| | | | | | | WorkingSetCommon::prepareForSnapshotChange shouldn't ever throw an exception, so it should be removed from the try/catch blocks in the delete and update stages. | ||||
* | SERVER-22178 Always retry sorted findAndModify upon write conflict. | Charlie Swanson | 2016-03-25 | 1 | -56/+69 |
| | | | | | | | | | Previously, if there was a WriteConflictException while actually doing the update or delete, we would retry the findAndModify, but if the update or delete stage detected that the document was already deleted or that it no longer matched the predicate, it would not retry. This patch ensures the findAndModify will be retried in either of those scenarios. | ||||
* | SERVER-22793 clear buffered WorkingSetIDs on storage engines without ↵ | David Storch | 2016-02-24 | 1 | -5/+1 |
| | | | | | | | document-level concurrency This ensures that the set of WorkingSetIDs does not grow without bound on MMAPv1. | ||||
* | SERVER-22531 use uassert instead of massert when stepping down during multi ↵ | Eric Milkie | 2016-02-09 | 1 | -1/+1 |
| | | | | delete. | ||||
* | SERVER-18826 Rename WorkingSet State Names from LOC to RID | James Wahlin | 2016-02-05 | 1 | -6/+6 |
| | | | | | | Also renamed: * WorkingSetMember::hasLoc() -> WorkingSetMember::hasRecordId * WorkingSetMember::loc -> WorkingSetMember::recordId | ||||
* | SERVER-16097 invariant(_collection) in DeleteStage::doRestoreState(). | Max Hirschhorn | 2016-02-04 | 1 | -0/+1 |
| | | | | A plan stage should only yield if the collection exists. | ||||
* | SERVER-22002 Do not retry findAndModify on MMAPv1 | Charlie Swanson | 2016-01-21 | 1 | -10/+0 |
| | |||||
* | SERVER-19377 PlanStage: make work() non-virt, add virt method doWork() | Jason Rassi | 2016-01-14 | 1 | -16/+1 |
| | | | | | Changes PlanStage::work() to be non-virtual. PlanStage::work() now updates _commonStats and calls new pure virtual method doWork(). | ||||
* | Revert "SERVER-22052 refactor Collection::deleteDocument to pass in the ↵ | Dianna Hohensee | 2016-01-06 | 1 | -1/+1 |
| | | | | | | to-be-deleted document directly as a parameter." This reverts commit 3b29ba46f2d7ce938ea0f084b2d0cb46d46c651e. | ||||
* | SERVER-22052 refactor Collection::deleteDocument to pass in the ↵ | Dianna Hohensee | 2016-01-05 | 1 | -1/+1 |
| | | | | to-be-deleted document directly as a parameter. | ||||
* | SERVER-21678 setting fromMigrate flag for deletes in oplog | Dianna Hohensee | 2016-01-04 | 1 | -1/+1 |
| | |||||
* | SERVER-21434 Retry a findAndModify if it conflicts on an update or delete | Charlie Swanson | 2015-11-13 | 1 | -2/+19 |
| | |||||
* | SERVER-19334: Refactor handling of status in delete stage | Geert Bosch | 2015-10-16 | 1 | -116/+129 |
| | |||||
* | SERVER-20111 Plan summary should only include the winning plan | Charlie Swanson | 2015-10-08 | 1 | -1/+1 |
| | |||||
* | SERVER-20512 Remove invalid calls to makeObjOwnedIfNeeded(). | Max Hirschhorn | 2015-10-07 | 1 | -11/+4 |
| | | | | | | | | | Don't try and make the object owned in DeleteStage if Collection::deleteDocument() throws a WriteConflictException because we would have already called saveState() on the child stage, which could have freed the memory underlying the WorkingSetMember. Similarly for UpdateStage with Collection::updateDocument(). | ||||
* | SERVER-19957 Transition to OWNED_OBJ state after performing the delete. | Max Hirschhorn | 2015-09-08 | 1 | -3/+9 |
| | | | | | | | Otherwise if a WriteConflictException were to occur during Collection::deleteDocument(), then a findAndModify operation with remove=true would not retry deleting that document because the associated WorkingSetMember no longer has a loc. | ||||
* | SERVER-19708: Move RecordCursor::seekExact to derived SeekableRecordStore class | Geert Bosch | 2015-08-28 | 1 | -1/+1 |
| | |||||
* | SERVER-16444 Only make BSONObj underlying WSM owned if not mmapv1. | Max Hirschhorn | 2015-08-27 | 1 | -2/+5 |
| | | | | | | | | | | Resolves the performance regression with mmapv1 introduced by 764e0c4 where documents were being copied in PlanStages that perform invalidations. Partially reverted changes to DeleteStage and UpdateStage because the document to return is not guaranteed to be owned after makeObjOwnedIfNeeded() is called when running with mmapv1. | ||||
* | SERVER-16444 Copy data in the query subsystem as needed. | Max Hirschhorn | 2015-08-24 | 1 | -5/+11 |
| | | | | | | | | | | | A WorkingSetMember in the LOC_AND_OBJ state must be made owned when: 1. Its WorkingSetID is cached across multiple calls to work(). 2. Multiple calls to next(), seekExact(), saveState(), etc. are performed on the same WiredTiger cursor in a single work() call. No longer necessary to always copy data out of WiredTiger buffers. | ||||
* | SERVER-19838 Remove member DeleteStageParams::shouldCallLogOp | Jason Rassi | 2015-08-10 | 1 | -12/+2 |
| | |||||
* | SERVER-19364 move query stage OperationContext pointer management into the ↵ | Qingyang Chen | 2015-08-04 | 1 | -11/+6 |
| | | | | base class | ||||
* | SERVER-17364 Don't stash RecoveryUnits across getMores | Mathias Stearn | 2015-07-16 | 1 | -3/+6 |
| | | | | | | | We now tell PlanExecutors to detach from their OperationContexts and to shed all storage engine resources before stashing the ClientCursor. This is a heavier weight operation than a normal save/restoreState which is no longer allowed to change the OperationContext. | ||||
* | SERVER-17364 Unify handling of child stages into PlanStage base class | Mathias Stearn | 2015-07-16 | 1 | -37/+11 |
| | | | | This is prep for adding more methods that need to propagate to children. | ||||
* | SERVER-18961 avoid iterating the entire working set when preparing for a ↵ | David Storch | 2015-07-01 | 1 | -5/+5 |
| | | | | | | | WiredTiger snapshot change Improves performance for query plans with a blocking stage when using the WiredTiger storage engine. In particular, full text search and geoNear queries should benefit. | ||||
* | SERVER-16889.5 PlanExecutor::getStats() and PlanStage::getStats() return ↵ | Qingyang Chen | 2015-06-29 | 1 | -5/+7 |
| | | | | unique_ptr | ||||
* | SERVER-16889 Modernize getExecutor*(), PlanExecutor::make() signatures | Qingyang Chen | 2015-06-26 | 1 | -4/+4 |
| | |||||
* | SERVER-18579: Clang-Format - reformat code, no comment reflow | Mark Benvenuto | 2015-06-20 | 1 | -232/+229 |
| | |||||
* | SERVER-17309 Replace std::auto_ptr<T> with std::unique_ptr<T> | Andrew Morrow | 2015-06-10 | 1 | -2/+2 |
| | |||||
* | SERVER-18807 modified ReplicationCoordinator::canAcceptWritesForDatabases to ↵ | Benety Goh | 2015-06-09 | 1 | -1/+1 |
| | | | | accept a namespace instead of a database name | ||||
* | SERVER-16444 New API for navigating RecordStores | Mathias Stearn | 2015-06-09 | 1 | -1/+3 |
| | |||||
* | SERVER-2454 Improve PlanExecutor::DEAD handling | James Wahlin | 2015-06-09 | 1 | -2/+1 |
| | |||||
* | SERVER-17633 add const qualifiers to PlanStage and Explain interfaces | David Storch | 2015-04-17 | 1 | -2/+2 |
| | |||||
* | SERVER-16063 Rewrite the findAndModify command. | Charlie | 2015-04-14 | 1 | -5/+52 |
| | | | | | | | | Changed UpdateStage to return the prior or newly-updated version of a document if request. also changed DeleteStage to return the deleted document if requested. Added explain support to the findAndModify command. | ||||
* | SERVER-17573 move OpObserver calls into the datalayer as much as possible ↵ | matt dannenberg | 2015-04-09 | 1 | -14/+0 |
| | | | | and eliminate repl bools | ||||
* | SERVER-17878 Rename GlobalEnvironmentExperiment to ServiceContext. | Andy Schwerin | 2015-04-03 | 1 | -2/+2 |
| | |||||
* | SERVER-13896 Replace logOp() with a more operation-aware observer interface | matt dannenberg | 2015-03-11 | 1 | -4/+6 |
| | |||||
* | SERVER-17062 rename NEED_FETCH to NEED_YIELD | David Storch | 2015-02-16 | 1 | -6/+6 |
| | |||||
* | SERVER-17062 Make query execution handle WriteConflictExceptions where possible | Mathias Stearn | 2015-02-13 | 1 | -41/+74 |
| | |||||
* | SERVER-17132 use snapshot id in DeleteStage, and re-check predicate | David Storch | 2015-02-03 | 1 | -3/+12 |
| | |||||
* | SERVER-17132: Added SnapshotId and Snapshotted and use them in query to make ↵ | Eliot Horowitz | 2015-02-02 | 1 | -2/+1 |
| | | | | sure we use correct versions of docs | ||||
* | SERVER-17132 Update and Delete stages need to forceFetchAllLocs after saveState | Mathias Stearn | 2015-02-02 | 1 | -0/+5 |
| | |||||
* | SERVER-16807 update and delete stages should skip over invalidated documents ↵ | David Storch | 2015-01-15 | 1 | -4/+8 |
| | | | | (MMAP v1) | ||||
* | SERVER-13256 Add usings and qualifications for names from namespace std | Andrew Morrow | 2015-01-15 | 1 | -0/+3 |
| | |||||
* | SERVER-16764 Make ReplicationCoordinator file names match the type names | Spencer T Brody | 2015-01-12 | 1 | -1/+1 |
| | |||||
* | SERVER-16675 force fetch RecordIds buffered by the query system on saveState() | David Storch | 2015-01-07 | 1 | -0/+12 |
| | | | | This fixes an issue with WiredTiger query isolation. | ||||
* | SERVER-16101 replace DeleteExecutor with ParsedDelete | David Storch | 2014-12-04 | 1 | -0/+10 |
| |