Details
-
Bug
-
Status: Open (View Workflow)
-
Major
-
Resolution: Unresolved
-
10.11, 11.4, 11.8
-
None
-
Can result in hang or crash
Description
binlog.binlog_old_versions fails in buildbot on release branches 10.11, 11.4 and 11.8, on the "==== Read binlog with v2 row events ====" step.
This is the same signature as MDEV-30945 (Closed/Fixed, 2023), which named this exact test and step as its earliest known occurrence: a MemorySanitizer use-of-uninitialized-value in bitmap_intersect (mysys/my_bitmap.c:395), reached via Rows_log_event::do_apply_event while replaying a v2 row-events binlog through mariadb-binlog / mysql_client_binlog_statement. That fix apparently did not hold, or a related code path regressed the same underlying uninitialized-bitmap issue.
Build Run Info
| Branch | Builder name | Date | Commit | Variant |
|---|---|---|---|---|
| 10.11 | amd64-msan-clang-22 | 2026-09-21T13:58:29Z | bdff5b8f2455acd226d242e16e8c2107f741d3da | mix |
| 11.4 | amd64-msan-clang-22 | 2026-09-21T17:13:43Z | 6c951fea5f62e1eb0ed7483ad8aa2cdcbfff5dd1 | mix |
| 11.8 | amd64-msan-clang-22 | 2026-09-18T19:03:53Z | ccf21ce23870f8122ee4cb1f184453b893797a78 | mix |
Failure text
==== Read binlog with v2 row events ====
|
Identical across all matched rows in the window.
Analysis
All 15 matching failures in the window (across 10.11, 11.4 and 11.8) occur on a single builder, amd64-msan-clang-22. That points at this specific MSAN test/build path rather than a broad per-branch regression - consistent with MDEV-30945's original signature (MSAN-only use-of-uninitialized-value in bitmap_intersect). The recurrence across three current release branches suggests the original fix either did not fully cover this replay path, or an unrelated change reintroduced the same class of uninitialized read.
See MDEV-30945 for the original fix and stack trace.
Created via the mtr-failure-bug skill.
Attachments
Issue Links
- relates to
-
MDEV-40415 Address replication MTR test failures
-
- Open
-