data corruption on arm64 and ppc64el
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
MySQL Server |
Unknown
|
Unknown
|
|||
mariadb-10.0 (Debian) |
Fix Released
|
Unknown
|
|||
mariadb-10.0 (Ubuntu) |
Fix Released
|
Undecided
|
Unassigned | ||
Trusty |
Invalid
|
Undecided
|
Unassigned | ||
Utopic |
Invalid
|
Undecided
|
Unassigned | ||
Vivid |
Fix Released
|
Undecided
|
Unassigned | ||
Wily |
Fix Released
|
Undecided
|
Unassigned | ||
mariadb-5.5 (Ubuntu) |
Invalid
|
Undecided
|
Unassigned | ||
Trusty |
Fix Released
|
Critical
|
Unassigned | ||
Utopic |
Fix Released
|
Undecided
|
Unassigned | ||
mysql-5.5 (Ubuntu) |
Invalid
|
Undecided
|
Unassigned | ||
Trusty |
Fix Released
|
Critical
|
Unassigned | ||
Utopic |
Fix Released
|
Critical
|
Unassigned | ||
mysql-5.6 (Ubuntu) |
Fix Released
|
Critical
|
Unassigned | ||
Trusty |
Fix Released
|
Critical
|
Unassigned | ||
Utopic |
Won't Fix
|
Critical
|
Unassigned | ||
Vivid |
Won't Fix
|
Critical
|
Unassigned | ||
Wily |
Fix Released
|
Critical
|
Unassigned |
Bug Description
I've verified that the following MariaDB issue also impacts MySQL:
https:/
[Impact]
Data corruption on arm64 and ppc64el
[Test Case]
Reproducibility varies across SoCs but, at least on one SoC, it reliably fails within seconds with these sysbench commands:
sysbench --num-threads=128 --max-requests=0 --max-time=1000 --test=oltp prepare
sysbench --num-threads=128 --max-requests=0 --max-time=1000 --test=oltp run
On failure, mysql will crash and sysbench will report a connection loss to the database.
[Regression Risk]
The fix touch some low level locking code, so there is always the risk that it introduces a problem on other platforms. However, this fix has been included in MariaDB since the 10.0.13 release in Aug, 2014 so it has had some real world testing.
tags: | added: patch |
Changed in mariadb-5.5 (Ubuntu Trusty): | |
importance: | Undecided → Critical |
status: | New → Confirmed |
Changed in mysql-5.6 (Ubuntu Vivid): | |
assignee: | nobody → recep (rbasak) |
Changed in mysql-5.6 (Ubuntu Vivid): | |
assignee: | recep (rbasak) → Robie Basak (racb) |
no longer affects: | mariadb-5.5 (Ubuntu Vivid) |
no longer affects: | mariadb-5.5 (Ubuntu Wily) |
Changed in mariadb-10.0 (Debian): | |
status: | Unknown → New |
Changed in mysql-5.6 (Ubuntu Vivid): | |
status: | Confirmed → In Progress |
Changed in mysql-5.6 (Ubuntu Utopic): | |
status: | New → In Progress |
Changed in mysql-5.6 (Ubuntu Trusty): | |
status: | Confirmed → In Progress |
Changed in mariadb-5.5 (Ubuntu Utopic): | |
status: | New → In Progress |
status: | In Progress → New |
Changed in mysql-5.5 (Ubuntu Utopic): | |
status: | New → In Progress |
description: | updated |
summary: |
- data corruption on arm64 + data corruption on arm64 and ppc64el |
Changed in mysql-5.5 (Ubuntu Trusty): | |
status: | New → In Progress |
Changed in mysql-5.6 (Ubuntu Utopic): | |
importance: | Undecided → Critical |
Changed in mysql-5.5 (Ubuntu Trusty): | |
importance: | Undecided → Critical |
Changed in mysql-5.5 (Ubuntu Utopic): | |
importance: | Undecided → Critical |
Changed in mariadb-10.0 (Debian): | |
status: | New → Fix Released |
Changed in mariadb-5.5 (Ubuntu): | |
status: | New → Fix Released |
Changed in mysql-5.5 (Ubuntu): | |
status: | New → Invalid |
Changed in mariadb-5.5 (Ubuntu): | |
status: | Fix Released → Invalid |
Apparently the fix I've proposed here caused regressions in MariaDB and was reworked: /mariadb. atlassian. net/browse/ MDEV-7026 /mariadb. atlassian. net/browse/ MDEV-7658
https:/
Unfortunately, the reworking reintroduces the problem on arm64:
https:/