summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Cut release 2.1.12.1.1Alex Gorrod2014-03-044-5/+31
|
* Fix wtperf make check, and versions in upgrading documentation.Alex Gorrod2014-03-042-8/+9
|
* Merge pull request #889 from wiredtiger/python-flushMichael Cahill2014-03-041-7/+18
|\ | | | | In python, flush during callbacks to write to sys streams. refs #883.
| * In python, flush during callbacks to wrute to sys streams. refs #883.Don Anderson2014-03-031-7/+18
|/
* Have the eviction thread update the oldest transaction ID in the system ↵Michael Cahill2014-03-031-0/+8
| | | | before each walk: we now use that to filter LRU candidates, and if a long-running transaction such as a checkpoint blocks all other threads in the system, the oldest ID has to be updated after the long-running transaction completes.
* Read threads should not bump the cache read generation: eviction is now ↵Michael Cahill2014-03-031-1/+1
| | | | responsible for that. Somehow this part of 30b1a822b26ff0531a0d31aa2895bc98b6f53d88 got lost.
* Use one check for whether to try forced eviction in a given session. Skip ↵Michael Cahill2014-03-033-32/+36
| | | | eviction of clean pages with modifications that are too recent, unless eviction gets aggressive.
* Don't use memory after we've freed it.Keith Bostic2014-03-011-4/+8
|
* Remove the assertion that every chunk being merged has a Bloom filter: under ↵Michael Cahill2014-03-011-4/+0
| | | | load, creating a Bloom filter can fail and retry, so we can't guarantee that all Bloom filters are created in order without more effort.
* Loosen the assertions around LSM merges and Bloom filters: it's okay to ↵Michael Cahill2014-02-281-4/+4
| | | | merge chunks in generations greater than zero even if they don't have Bloom filters.
* Merge pull request #888 from wiredtiger/rec-skip-cleanAlex Gorrod2014-02-284-27/+31
|\ | | | | Fix a bug regarding when clean pages can be evicted
| * Add comments explaining the reconciliation transactional read logic, ↵Michael Cahill2014-02-281-2/+15
| | | | | | | | including when pages become clean.
| * Rename disk_txn to reflect what we're actually using it for.Michael Cahill2014-02-283-4/+4
| |
| * Fix another bug regarding when pages can be evicted: have reconcilition ↵Michael Cahill2014-02-282-21/+12
|/ | | | | | track the maximum transaction ID on the page (regardless of whether it is skipped). A page can only be evicted when it is clean and has no updates that are too new. refs #884
* Merge pull request #887 from wiredtiger/merge-bloom-fixAlex Gorrod2014-02-282-9/+18
|\ | | | | Make sure all chunks involved in a merge have Bloom filters
| * Switch || to && to fix build.Alex Gorrod2014-02-281-1/+1
| |
| * Have LSM compact logic match regular merges exactly: always create Bloom ↵Michael Cahill2014-02-282-5/+3
| | | | | | | | filters first. If there are any concurrent reads in the tree, they will benefit from the Bloom filters, and this avoids having different paths through the merge code.
| * Make sure all chunk involved in a merge have Bloom filters. Fix a race ↵Michael Cahill2014-02-282-5/+16
|/ | | | | | where a chunk is merged and removed from the tree during creation of a Bloom filter. refs #809
* Display the failed open path, it makes debugging easier.Keith Bostic2014-02-271-3/+2
|
* monitor_set might have been used without being initialized.Keith Bostic2014-02-272-29/+28
| | | | otherwise, lint, whitespace, style
* Merge pull request #847 from wiredtiger/checkpoint-directioAlex Gorrod2014-02-2712-17/+40
|\ | | | | Do direct I/O reads on checkpoints to reduce the VM pressure of LSM workloads
| * Update documentation for direct_io configuration.Alex Gorrod2014-02-272-3/+12
| |
| * Revert unnecessary test changes.Alex Gorrod2014-02-272-13/+4
| |
| * Don't automatically enable direct I/O on checkpoints.Alex Gorrod2014-02-274-8/+5
| | | | | | | | | | Using direct I/O requires careful selection of page sizes, and doesn't always improve performance. It's not worth enabling it automatically.
| * Update test/salvage to configure the right buffer size for direct io.Alex Gorrod2014-02-271-4/+9
| |
| * Do buffer alignment checks when reading from checkpoints.Alex Gorrod2014-02-272-1/+9
| | | | | | | | Update test_checkpoint01 to use valid buffer sizes.
| * Merge branch 'develop' into checkpoint-directioMichael Cahill2014-02-2653-731/+1615
| |\
| * \ Merge branch 'develop' into checkpoint-directioMichael Cahill2014-02-0775-1724/+2658
| |\ \
| * | | Do direct I/O reads on checkpoints to reduce the VM pressure of LSM workloads.Michael Cahill2014-01-1711-16/+29
| | | |
* | | | Add compression to voxer configs.Susan LoVerso2014-02-262-0/+2
| | | |
* | | | Change pages_walked to uint64_t.Susan LoVerso2014-02-261-2/+3
| | | |
* | | | Merge pull request #877 from wiredtiger/lsm-fixessueloverso2014-02-2612-75/+143
|\ \ \ \ | | | | | | | | | | Improve LSM performance with wtperf tests 1-4
| * \ \ \ Merge branch 'develop' into lsm-fixesSusan LoVerso2014-02-265-32/+51
| |\ \ \ \ | |/ / / / |/| | | |
* | | | | Don't create an illegal config string.Susan LoVerso2014-02-261-1/+1
| | | | |
* | | | | Fix bug finding table count with multiple tables.Susan LoVerso2014-02-261-4/+9
| | | | |
* | | | | Add some comments and fix some style.Susan LoVerso2014-02-262-17/+21
| | | | |
* | | | | Coverity fixes.Susan LoVerso2014-02-261-4/+10
| | | | |
* | | | | Allow overlapping strings when setting a buffer's contents, it's notKeith Bostic2014-02-261-1/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | insane __wt_buf_set might be used to shuffle strings in an existing WT_ITEM buffer. Reference #881, but this doesn't close it, we still want to understand why it can happen.
* | | | | Fix typo when closing config_parser handle, complain if theKeith Bostic2014-02-261-5/+7
| | | | | | | | | | | | | | | | | | | | config_parser.close method errors.
* | | | | Fix build warning in wtperf.Alex Gorrod2014-02-261-0/+1
| |_|_|/ |/| | |
| * | | Merge branch 'develop' into lsm-fixesMichael Cahill2014-02-2628-637/+970
| |\ \ \ | |/ / / |/| | |
* | | | Merge pull request #873 from wiredtiger/config-parse-apiMichael Cahill2014-02-2627-516/+842
|\ \ \ \ | | | | | | | | | | Add configuration parsing to the public API.
| * \ \ \ Merge pull request #875 from wiredtiger/wtperf-multiple-databasesAlex Gorrod2014-02-265-149/+291
| |\ \ \ \ | | | | | | | | | | | | Add ability for wtperf to run with multiple databases in different threa...
| | * | | | Fixes to wtperf as per review comments.Alex Gorrod2014-02-262-36/+27
| | | | | |
| | * | | | Merge branch 'config-parse-api' into wtperf-multiple-databasesAlex Gorrod2014-02-2635-538/+623
| | |\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Conflicts: bench/wtperf/wtperf.c bench/wtperf/wtperf.h
| | * \ \ \ \ Merge branch 'config-parse-api' into wtperf-multiple-databasesAlex Gorrod2014-02-142-10/+13
| | |\ \ \ \ \
| | * | | | | | Add ability for wtperf to run with multiple databases in different threads.Alex Gorrod2014-02-145-148/+299
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This can be handy when testing shared cache behavior or generating different I/O workloads.
| * | | | | | | Add upgrade documentation about configuration string parsing changes.Alex Gorrod2014-02-262-0/+15
| | | | | | | |
| * | | | | | | Update documentation for wiredtiger_config_parser_openAlex Gorrod2014-02-261-2/+2
| | | | | | | |
| * | | | | | | Fix error handling in wtperf config parsing.Alex Gorrod2014-02-261-6/+6
| | |_|/ / / / | |/| | | | |
View Lorry Controller Status