| Commit message (Expand) | Author | Age | Files | Lines |
|---|
| * | WT-2251 Clear the original ref->addr before an in-memory split to avoid use-a... | Michael Cahill | 2015-12-01 | 1 | -0/+11 |
| * | Merge branch 'develop' into WT-2251-ref-addr-leak | Michael Cahill | 2015-12-01 | 1 | -8/+11 |
| |\ |
|
| | * | SERVER-21691 Avoid blocking in-memory splits. | Michael Cahill | 2015-12-01 | 1 | -8/+11 |
| * | | WT-2251 Free addresses when we discard deleted page references. | Michael Cahill | 2015-12-01 | 1 | -0/+1 |
| |/ |
|
| * | Rename __wt_page_can_split to be __wt_leaf_page_can_split, we split | Keith Bostic | 2015-11-27 | 1 | -1/+1 |
| * | WT-2244 Don't wait for transactions after splits. | Michael Cahill | 2015-11-27 | 1 | -18/+0 |
| * | SERVER-21553 Fix a line removed by mistake: set the address during splits. | Michael Cahill | 2015-11-27 | 1 | -0/+1 |
| * | SERVER-21553 Enable fast-path truncate after splits. | Michael Cahill | 2015-11-27 | 1 | -3/+12 |
| * | Merge pull request #2324 from wiredtiger/wt-2230-multi-split-error-path | Alex Gorrod | 2015-11-26 | 1 | -4/+4 |
| |\ |
|
| | * | WT-2230: The split_multi_inmem_fail function calls wt_free_ref with | Keith Bostic | 2015-11-20 | 1 | -4/+4 |
| * | | WT-2237 Have threads publish unique transaction IDs so that updates always be... | Michael Cahill | 2015-11-24 | 1 | -2/+2 |
| |/ |
|
| * | SERVER-21568 Free page_del information in refs causing reverse splits. | Michael Cahill | 2015-11-20 | 1 | -1/+12 |
| * | SERVER-21568 Free deleted page structures as soon as possible. | Michael Cahill | 2015-11-20 | 1 | -11/+6 |
| * | SERVER-21568 Fix a use after-free. | Michael Cahill | 2015-11-20 | 1 | -10/+13 |
| * | WT-2224 Cleanup and fix. | Michael Cahill | 2015-11-18 | 1 | -23/+16 |
| * | WT-2224 Track which deleted refs are discarded by a split. | Michael Cahill | 2015-11-18 | 1 | -10/+23 |
| * | WT-2182: don't leave child refs locked; when we hit the result_entries | Keith Bostic | 2015-11-17 | 1 | -5/+14 |
| * | WT-2182: we can't let newly created internal pages split until we've | Keith Bostic | 2015-11-13 | 1 | -42/+77 |
| * | Fix a comment. | Keith Bostic | 2015-11-13 | 1 | -1/+1 |
| * | Use a better name for the "no-evict" function. | Keith Bostic | 2015-11-13 | 1 | -4/+4 |
| * | I think there's a race in setting the mod_split_txn field as each new | Keith Bostic | 2015-11-13 | 1 | -16/+36 |
| * | WT-2182: factor out the code to clean up moved WT_REF structures, after | Keith Bostic | 2015-11-13 | 1 | -103/+79 |
| * | Merge branch 'wt-2182-split-up-the-tree-2' into wt-2182-keith | Keith Bostic | 2015-11-13 | 1 | -1/+1 |
| |\ |
|
| | * | whitespace | Keith Bostic | 2015-11-13 | 1 | -1/+1 |
| * | | Make the fix-up loop in __split_internal match the main loop that creates | Keith Bostic | 2015-11-13 | 1 | -2/+2 |
| |/ |
|
| * | And, remove the test code. | Keith Bostic | 2015-11-13 | 1 | -3/+3 |
| * | My mistake, ref->page will be NULL during eviction triggered reverse splits, | Keith Bostic | 2015-11-13 | 1 | -5/+9 |
| * | spelling typo | Keith Bostic | 2015-11-13 | 1 | -1/+1 |
| * | Rework 2a71849, some new verbose messages. | Keith Bostic | 2015-11-13 | 1 | -17/+15 |
| * | WT-2182 Add more verbose tracking to splits | Alex Gorrod | 2015-11-13 | 1 | -0/+15 |
| * | Fix up a couple of comments, minor KNF, minor re-work so the split-root and | Keith Bostic | 2015-11-12 | 1 | -16/+13 |
| * | Sanity check both split-parent and split-root the same way, there has | Keith Bostic | 2015-11-12 | 1 | -17/+22 |
| * | Remove WT_PAGE.{pg_intl_deepen_split_append,pg_intl_deepen_split_last}, | Keith Bostic | 2015-11-12 | 1 | -24/+2 |
| * | Merge branch 'wt-2182-split-up-the-tree-2' into wt-2182-split-simplify | Alex Gorrod | 2015-11-12 | 1 | -22/+31 |
| |\ |
|
| | * | change split verbose messages to distinguish between root and internal | Keith Bostic | 2015-11-11 | 1 | -2/+4 |
| | * | Merge branch 'develop' into wt-2182-split-up-the-tree-2 | Keith Bostic | 2015-11-09 | 1 | -20/+27 |
| | |\ |
|
| | | * | The last change was too simplistic: we need to set WT_PAGE_UPDATE_IGNORE | Keith Bostic | 2015-11-06 | 1 | -24/+27 |
| | | * | Release the allocated page on error when creating a new in-memory page | Keith Bostic | 2015-11-06 | 1 | -0/+4 |
| * | | | WT-2182 Simplify the split deepen logic. | Alex Gorrod | 2015-11-10 | 1 | -60/+17 |
| |/ / |
|
| * | | When splitting a page, the parent may not have yet been dirtied, | Keith Bostic | 2015-11-09 | 1 | -0/+9 |
| * | | If we replace a page index, mark the page dirty. | Keith Bostic | 2015-11-09 | 1 | -3/+6 |
| * | | Revert a7ac7c6, don't acquire the split-generation until after the actual | Keith Bostic | 2015-11-08 | 1 | -12/+7 |
| * | | If a reverse split fails, let __evict_delete_ref() unlock the child, | Keith Bostic | 2015-11-08 | 1 | -8/+1 |
| * | | We should be able to detect an internal page split race by noticing we're | Keith Bostic | 2015-11-05 | 1 | -0/+6 |
| * | | Acquire a new split generation at the start of each split routine, it's | Keith Bostic | 2015-11-05 | 1 | -3/+9 |
| * | | Remove code that coped with copying the original page's WT_REF, we no | Keith Bostic | 2015-11-05 | 1 | -3/+0 |
| * | | Consistently update the statistics at the start of the function. | Keith Bostic | 2015-11-04 | 1 | -6/+6 |
| * | | Now we're only deepening at the root, the name is "root_refp", not | Keith Bostic | 2015-11-04 | 1 | -9/+9 |
| * | | Now we're only deepening at the root, the name is "root", not "parent"; | Keith Bostic | 2015-11-04 | 1 | -41/+40 |
| * | | Rename most of the __split_ref_move arguments, that function is no | Keith Bostic | 2015-11-04 | 1 | -33/+46 |
