|
Ran
|
Jobs
1
|
Files
479
|
Run time
1min
|
Badge
README BADGES
|
push
github
box: set TXN_WAIT_SYNC with xrow rather than limbo state during recovery During recovery we set the `TXN_WAIT_SYNC` flag based on the limbo's state rather than respecting the xrow's flags. This does not play well with fully local transactions that were committed using the `TXN_FORCE_ASYNC` flag: during runtime they bypass the limbo as expected, but during recovery we do not set this flag, and they are appended to the limbo, if it is not empty, even though they are not supposed to be, since their xrows does not have the `wait_sync` flag. To fix this, during recovery let's set all the required flags related to synchronous replication while processing the xrow, and ignore the limbo emptiness, if we are in recovery and processing local transactions (i.e., not applying remote transactions). Closes #11053 NO_CHANGELOG=<this version is not affected by backported bugfix> NO_DOC=<this version is not affected by backported bugfix> NO_TEST=<this version is not affected by backported bugfix> (cherry picked from commit 4ba187909)
63544 of 114902 branches covered (55.3%)
4 of 4 new or added lines in 1 file covered. (100.0%)
37 existing lines in 12 files now uncovered.94468 of 109740 relevant lines covered (86.08%)
1907966.89 hits per line
| Lines | Coverage | ∆ | File |
|---|---|---|---|
| 1 |
94.54 |
0.66% | src/box/replication.cc |
| 1 |
92.35 |
-0.17% | src/box/txn_limbo.c |
| 1 |
95.68 |
-0.72% | src/box/vy_point_lookup.c |
| 1 |
80.0 |
-2.05% | src/lib/core/clock.c |
| 2 |
97.96 |
0.0% | src/box/txn.h |
| 2 |
92.09 |
0.0% | src/box/vinyl.c |
| 3 |
90.01 |
-0.16% | src/box/applier.cc |
| 3 |
93.31 |
-1.0% | src/box/gc.c |
| 3 |
89.36 |
-0.32% | src/box/vy_scheduler.c |
| 4 |
89.44 |
-0.41% | src/box/memtx_engine.cc |
| 4 |
93.31 |
-1.27% | src/lib/swim/swim_io.c |
| 12 |
92.72 |
-0.06% | src/box/txn.c |
| ID | Job ID | Ran | Files | Coverage | |
|---|---|---|---|---|---|
| 1 | 14490543570.1 | 479 |
86.08 |
GitHub Action Run |
| Coverage | ∆ | File | Lines | Relevant | Covered | Missed | Hits/Line | Branch Hits | Branch Misses |
|---|