summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Cut release 1.3.4.1.3.4Michael Cahill2012-10-197-23/+120
|
* Use the O_NOATIME flag: it may improve performance on some systems.Michael Cahill2012-10-191-0/+5
|
* lsm_cursor.c:716:3: error: 'smaller' may be used uninitialized in this functionMichael Cahill2012-10-191-2/+4
| | | | lsm_cursor.c:718:3: error: 'larger' may be used uninitialized in this function
* Reset any old cursor position before an LSM search.Michael Cahill2012-10-191-3/+40
| | | | Long term, we need a more general approach to cleaning up the old LSM cursor state, but this change is important in that it keeps the number of active hazard references in a searching thread less than or equal to 1.
* Add an API to pass a config string to __wt_bloom_create.Michael Cahill2012-10-198-28/+49
|
* With the updated scenarios, test_schema03 needs far few file descriptors.Michael Cahill2012-10-191-1/+1
| | | | (This changes allows test_schema03 to be run as part of normal automated testing).
* Merge pull request #369 from wiredtiger/lsm-schemaagorrod2012-10-1829-492/+621
|\ | | | | Support LSM as a backing source for schema-level objects (tables, column groups and indices)
| * Get test_schema03 running with LSM.Michael Cahill2012-10-191-14/+16
| | | | | | | | Significantly crank down the size of the schemas it generates: it is better to have it running consistently with moderate-sized scenarios than requiring so many resources that we avoid running it in normal testing.
| * Fix LSM index searches.Michael Cahill2012-10-192-25/+42
| | | | | | | | | | | | | | | | | | | | The main issue was LSM search_near was not always returning the closest key to the search key, which calling code expects. It now tries hard to find the smallest cursor larger than the search key, and only if no larger record exists does it return the largest record smaller than the search key. While in the area, clean up the index search code and make it slightly more efficient.
| * Test LSM indices in test_schema02.Michael Cahill2012-10-191-10/+25
| |
| * Support indices on LSM by using 1-byte (padding) values, until LSM supports ↵Michael Cahill2012-10-194-1/+17
| | | | | | | | empty values natively.
| * Push the "owning" cursor through the LSM open_cursor function, so schema ↵Michael Cahill2012-10-196-7/+12
| | | | | | | | objects are closed in the correct order.
| * Run some standard tests against table-on-LSM.Michael Cahill2012-10-184-13/+24
| |
| * Calculate LSM data source names for column groups and indices.Michael Cahill2012-10-184-16/+45
| |
| * Handle NULL config stacks in LSM cursor open.Michael Cahill2012-10-181-1/+1
| |
| * lintMichael Cahill2012-10-182-4/+5
| |
| * Change direct calls to __wt_curfile_open to generic __wt_open_cursor.Michael Cahill2012-10-187-28/+18
| |
| * Apply operations to column group and index data sources.Michael Cahill2012-10-1810-266/+240
| | | | | | | | Change index cursors to be a generic wrapper around the data source cursor, rather than always extending a file cursor.
| * Merge branch 'develop' into lsm-schemaMichael Cahill2012-10-1861-590/+1199
| |\ | |/ |/| | | | | Conflicts: dist/api_data.py
* | Further cleanup of forced eviction comments and code.Michael Cahill2012-10-187-50/+18
| |
* | Fix a bug in LSM where aborted merges could cause subsequent opens to fail.Alex Gorrod2012-10-182-2/+17
| | | | | | | | | | Also fix a bug in LSM stats - where we returned an error for empty chunks that do not have a checkpoint.
* | another leakMichael Cahill2012-10-181-0/+2
| |
* | Don't force O_NOATIME on -- it may have unintended consequences, and on ↵Michael Cahill2012-10-181-3/+0
| | | | | | | | modern Linux systems shouldn't make a huge difference.
* | Have LSM merges sleep for much less than a second: this limits how quickly ↵Michael Cahill2012-10-181-3/+3
| | | | | | | | we can cycle through operations in test/fops.
* | Tweaks to file configuration for reads on Linux (disable readahead and ↵Michael Cahill2012-10-182-3/+22
| | | | | | | | access time updates).
* | Don't stop evicting until we reach the target, have eviction wake up ↵Michael Cahill2012-10-186-25/+27
| | | | | | | | periodically regardless of whether the application signals it. This latter requires a "timed condition wait" operation.
* | Wait for a while before looking for LSM major merges, in case merges catch ↵Michael Cahill2012-10-183-17/+36
| | | | | | | | up with inserts.
* | Add a macro to calculate the fixed-size slots per a byte count, noKeith Bostic2012-10-171-5/+7
| | | | | | | | real change, just encapsulation.
* | When doing a fast-delete, lock down the page before we unpack theKeith Bostic2012-10-171-24/+28
| | | | | | | | | | | | | | | | | | | | | | | | on-page cell -- while probably close to impossible, it's theoretically possible the page could be instantiated, modified and reconciled while we're sleeping, in which case the WT_REF.addr field would no longer point on-page. We can't set the page's state to WT_REF_DELETED if __wt_txn_modify_ref fails, that's the function that schedules rollback of the page state after the transaction fails, and if it fails we're probably not able to roll-back the state of this page.
* | whitespaceKeith Bostic2012-10-171-1/+1
| |
* | Replace a few strncmp() calls with WT_PREFIX_MATCH.Keith Bostic2012-10-171-11/+8
| |
* | Fix a comment.Keith Bostic2012-10-171-1/+1
| |
* | Drop the compaction limit, compact any file that's at least 10K.Keith Bostic2012-10-171-1/+1
| |
* | lint, don't ignore function return value.Keith Bostic2012-10-171-2/+2
| |
* | Add backup, config and statistics object types, that way we'llKeith Bostic2012-10-171-28/+18
| | | | | | | | | | report an error for "verify statistics:foo" instead of trying to verify a the URI "table:statistics:foo".
* | Compaction is also supported on objects of type colgroup and index.Keith Bostic2012-10-171-2/+3
| |
* | Limit WT_SESSION::compact to objects of type colgroup, file, index, lsmKeith Bostic2012-10-171-0/+9
| | | | | | | | and table, and simply return success for lsm objects.
* | Add lsm: type to the list of unsupported object types.Keith Bostic2012-10-171-0/+1
| |
* | Add a comment so we (hopefully) don't cut-and-paste the line into theKeith Bostic2012-10-173-0/+6
| | | | | | | | next command we add.
* | Make command names/descriptions line up, remove dumpfile it's no longerKeith Bostic2012-10-171-2/+1
| | | | | | | | a command.
* | Upgrade will have a progress report at some point, put in a place-holderKeith Bostic2012-10-171-1/+1
| | | | | | | | so the wt utility doesn't look wrong.
* | The rename call doesn't have any verbose flag behaviors, don't outputKeith Bostic2012-10-171-2/+0
| | | | | | | | an extra newline.
* | Change the loop to read in the extent blocks we need, and the loop toKeith Bostic2012-10-171-43/+55
| | | | | | | | | | process the checkpoints, look the same, the former was too complex to no advantage.
* | Fake checkpoints may have the delete flag set, ignore them when rollingKeith Bostic2012-10-171-0/+2
| | | | | | | | checkpoints forward. Reference #366.
* | typoKeith Bostic2012-10-171-1/+1
| |
* | Merge pull request #365 from wiredtiger/compactMichael Cahill2012-10-1731-359/+863
|\ \ | | | | | | | | | | | | On-line compaction. refs #248
| * \ Merge branch 'develop' into compactKeith Bostic2012-10-1710-11/+32
| |\ \ | |/ / |/| |
* | | ex_all.c: scoping typo.Michael Cahill2012-10-171-2/+2
| | |
* | | One more leak, not closing the connection in ex_all.cMichael Cahill2012-10-171-0/+2
| | |
* | | s/WildTiger/WiredTiger/Michael Cahill2012-10-174-4/+3
| | |
View Lorry Controller Status