Console View

Categories: connectors experimental galera main

Legend:

Passed

Failed

Warnings

Failed Again

Running

Exception

Offline

No data

connectors

experimental

galera

main

fcc5fc891660...

fcc5fc8916609849cc6397267fa118bb738ffb1a

Kristian Nielsen

knielsen@knielsen-hq.org

Merge 11.8 to 12.3

Signed-off-by: Kristian Nielsen <[email protected]>

fa723683b02b...

fa723683b02b450a1c221b9ac60993d8de004aa7

Raghunandan Bhat

raghunandan.bhat96@gmail.com

MDEV-39559: Assertion `b` failed in `my_strnncoll_xxxx_nopad_ci`

Problem:
When verifying a `UNIQUE` constraint over an empty BLOB/TEXT column,
`Field_blob::cmp(a_ptr, b_ptr)` extracts a NULL data pointer (with
length 0) from the record slot and forwards it to the charset-level
comparison, which asserts non-NULL pointers (added in MDEV-35717).

A zero-length blob is allowed to have a NULL data pointer in its
record slot; `Field_blob::val_decimal()` already treats (NULL, 0) as
an empty value. The comparison paths missed that substitution.

Fix:
- In `Field_blob::cmp(a_ptr, b_ptr)`and
`Field_blob::key_cmp(key_ptr, max_key_length)`, substitute "" for a
NULL data pointer before delegating to the comparison.
- Add a debug assert in each of the above functions to document the
invariant that a NULL pointer is only valid alongside zero length
and (NULL, length>0) is invalid.

eef3f3a93508...

eef3f3a935087c9ae74522035bf728b565a7db44

drrtuy

drrtuy@gmail.com

chore: update DuckDB docs and info.

eccdef627382...

eccdef627382b7b1f5338450ba328d16216c4fa2

Kristian Nielsen

knielsen@knielsen-hq.org

Post-merge fix

Signed-off-by: Kristian Nielsen <[email protected]>

ec80c9d7bad1...

ec80c9d7bad1b4818563654e8146f6e9833a4b33

Tony Chen

ochenton@amazon.com

MDEV-39828 Stabilize query_cache_executable_comments counter checks

The test asserted absolute Qcache_inserts/Qcache_hits values but reset them
with FLUSH STATUS which does not clear those global counters.
The expected values were therefore cumulative from a pristine server, so the
test drifted by a constant offset whenever the mysqld instance was reused with
prior query cache activity (fulltest server reuse, --repeat).

E.g:

```
./build/mysql-test/mysql-test-run.pl --parallel=1 --repeat=2 main.query_cache_executable_comments
```

Use FLUSH GLOBAL STATUS instead, matching every other query cache test
(query_cache.test, query_cache_with_views.test, subselect_cache.test).
This zeroes the SHOW Qcache counters, making each case independent
of server history.

All new code of the whole pull request, including one or several files
that are either new files or modified ones, are contributed under the
BSD-new license. I am contributing on behalf of my employer Amazon Web
Services, Inc.

ea06ef9f96fd...

ea06ef9f96fdb18c9402312056d02cee1e95e6f3

Kristian Nielsen

knielsen@knielsen-hq.org

Merge branch '12.3' into 13.0

e8fde9720646...

e8fde97206464833c2c82b08aa66ff649ddbe4f4

Yuchen Pei

ycp@mariadb.com

MDEV-24813 Signal full scan to storage engines.

When starting to do a full table/index scan without a WHERE or JOIN
condition, tell the storage engine so and the corresponding
ulong-truncated LIMIT.

Include an innodb implementation: added an innodb switch
table_lock_on_full_scan, so that when the switch is on, on receiving
the full scan signal from the sql layer, if the truncated LIMIT is
ULONG_MAX (likely no LIMIT), attempt to acquire a table lock.

Updated tests that have different results with the switch on.

(Comment and code edited by Sergei Petrunia)

MDEV-24813 Drop low-value table_lock test combinations

The three deadlock_*_race tests cannot reach their DEBUG_SYNC race
under a table lock (it degenerates to a timeout), and the three I_S
tests only restate a lock-mode change already covered
by innodb_full_scan.test.

- Addressed review comments

d9c364ec4dc9...

d9c364ec4dc967ffa7a02a030c5b2c1fa8faf8a7

Jan Lindström

jan.lindstrom@mariadb.com

MDEV-38870 : Galera test failure on galera.MDEV-38201

Test changes only. Moved wait condition where node should
disconnect from cluster because it has become inconsistent.
After that next FLUSH HOST based on timing could return
not supported or timeout.

cecd41266599...

cecd41266599a41d5c7daf177e2c345aa2c0afe5

Kristian Nielsen

knielsen@knielsen-hq.org

Merge branch '11.8' into 12.3

cbaa0fc1df69...

cbaa0fc1df69b971987ce9abfcca9b5c250ec813

Pekka Lampio

pekka.lampio@galeracluster.com

MDEV-38386 Fix incomplete cleanup in Galera MTR tests failing under --repeat

A number of Galera MTR tests pass on the first run but fail on a second
--repeat iteration, because server, cluster or filesystem state leaks
across runs and the test does not restore a clean starting state.

Fix the cleanup (or force a fresh cluster) in the affected tests. Each
fix was verified with --repeat=2 --force.

1. Stale async-slave GTID position (11 tests)

RESET SLAVE [ALL] does not clear gtid_slave_pos. As the master does
RESET MASTER in cleanup, on the next run the slave considers the
events already applied and skips them, so the replicated tables never
appear. Clear the position with SET GLOBAL gtid_slave_pos = "".

2. Leftover binlog GTID state from trailing cleanup (1 test)

Trailing DROP TABLE / mtr.add_suppression statements ran after the
.inc's reset master and re-populated node_2's binlog. gtid_binlog_state
keeps the latest seqno per (domain, server_id) pair, so a stray
0-2-<n> survived into the next run and broke the state comparison.
Reorder the cleanup and reset node_2's binlog last.

3. Cluster-global, one-time or time-window state (11 tests)

The wsrep GTID domain seqno is cluster-global and is not reset by
reset master (nor by a mid-test SST rejoin); error-log contents,
warning-flood suppression timers and one-time bootstrap behaviour are
likewise not restored by in-test cleanup. Force a fresh cluster with
include/force_restart.inc.

4. Leftover filesystem artifacts (1 test)

mariabackup refuses to back up into a non-empty target directory, so
the leftover target dirs from the previous run made the backup fail
silently and the expected log messages never appeared. Remove the
target directories in cleanup.

caafc23878e0...

caafc23878e010f11d602777e4e2ed812118fd57

Dave Gosselin

dave.gosselin@mariadb.com

MDEV-39207: mark test as not_embedded

Test fails on embedded CI because original not_embedded flag was not preserved

c79ca44f24ad...

c79ca44f24ad5cd8bb45a06e8ad91a73e4148c55

Daniel Black

daniel@mariadb.org

MDEV-31209 Queries with window functions do not obey KILL / max_statement_time

Window functions run in a loop in
Frame_cursor::compute_values_for_current_row which can include a large
number of rows.

Adjust this function to check for the current thd being killed by only
every 256 rows, so as not to destroy any CPU pipelining or similar.

c46441bc5f19...

c46441bc5f19f6b10dc1c3a2513d963c496e6578

Kristian Nielsen

knielsen@knielsen-hq.org

Post-merge fixes

Signed-off-by: Kristian Nielsen <[email protected]>

b9809c07b412...

b9809c07b412456b194ab342e26d6432e85edf09

Kristian Nielsen

knielsen@knielsen-hq.org

Document the correct use of replication terminology

Signed-off-by: Kristian Nielsen <[email protected]>

b88fe40601a2...

b88fe40601a252c29fde7432e04d9ed7e9897c40

Dave Gosselin

dave.gosselin@mariadb.com

MDEV-39952: Skip binlog_in_engine tests that need mariabackup

Several tests in the binlog_in_engine suite run the mariabackup
binary but did not check that it was built. When the server is
built with WITH_MARIABACKUP=OFF, the test runner leaves the
mariabackup command empty and these tests fail instead of being
skipped.

Add include/have_mariabackup_binary.inc, which skips the test when
the binary is not available, and source it from the affected tests.
Unlike have_mariabackup.inc, it does not also require socat or nc,
because these tests copy the backup to a local directory rather than
streaming it.

b27ccf7b01ff...

b27ccf7b01ff36cb85cfc45f1b823998d2d450fa

Andreas Schwab

schwab@suse.de

MDEV-39925: Fix error reporting in create_libaio

The io_setup function in libaio returns a negated errno value on error,
but strerror expects a normal errno value.

ae136494fb2a...

ae136494fb2a447d74c0e6cc747461f96e2ec17f

Fariha Shaikh

farihais@amazon.com

MDEV-39928 Fix GitLab CI centos9 job failure

The centos9 job uses yum-builddep -y mariadb-server to install build
dependencies, but the mariadb-server source package has been removed
from CentOS Stream 9 repositories. Replace with explicit installation of
the required build dependencies.

All new code of the whole pull request, including one or several files
that are either new files or modified ones, are contributed under the
BSD-new license. I am contributing on behalf of my employer Amazon Web
Services, Inc.

abaef32ac153...

abaef32ac1535fc2e430483653de3edce5ca3c75

Marko Mäkelä

marko.makela@mariadb.com

MDEV-39862 innodb_log_archive=ON corruption after modifying file parameters

log_t::set_buffered(), log_t::set_write_through(): If a log file switch
with innodb_log_archive=ON is in progress, ignore a
SET GLOBAL statement that would attempt to modify
innodb_log_file_buffering or innodb_log_file_write_through.

This prevents the log from becoming corrupted. The problem was that we
would close log_sys.log.m_file that pointed to the latest log file and
then reopen a second handle to the previous log file, which
log_sys.resize_log.m_file is already pointing to. As a result, log records
would be written to the wrong log file, causing the log to be corrupted.

The statements SET GLOBAL innodb_log_file_buffering and
SET GLOBAL innodb_log_file_write_through will also be ignored when a
SET GLOBAL innodb_log_file_size operation is in progress on a
circular-format log (ib_logfile0). The statements will have an effect
when InnoDB is holding only one log file open, which should be most
of the time. These statements have no effect when the log file is
mapped to persistent memory.

Whether a requested change took place can be checked by executing
a statement like the following:
SELECT @@GLOBAL.innodb_log_file_buffering;

Tested by: Saahil Alam

a7e2bfa2e25d...

a7e2bfa2e25d22c399f91603f244dab87fc1035d

Yuchen Pei

ycp@mariadb.com

[fixup] Move tests requiring an example plugin from sys_vars.session_track_system_variables_basic to a separate test

A follow up to 1f56d9c3feeeca82661cbe57cb628207c8b186f8. This restores
test coverage when the example plugin is not built.

a7a2bfda2a15...

a7a2bfda2a15b8b42693f3d99d8b27edb023fe31

Dave Gosselin

dave.gosselin@mariadb.com

MDEV-39952: Skip tests that need mariabackup

Skips tests that require mariabackup if mariabackup was not
built (WITH_MARIABACKUP=OFF).

Backport of the same MTR change from 12.3 but applied to
additional tests.

a344f1d708e0...

a344f1d708e052291601e71986b43ad561bdbbdf

Dave Gosselin

dave.gosselin@mariadb.com

MDEV-38669: ASAN fail in Item_func_collect::add during ST_COLLECT

st_collect read past the end of its input because it assumed
that the input was a valid geometry. Check that the length of the
input is at least long enough to possibly contain a valid geometry.

96ca4d3d5aa4...

96ca4d3d5aa416eb26c4de8bf3e681bb5185ae5e

Dave Gosselin

dave.gosselin@mariadb.com

MDEV-38158: Incorrect query result

When setup_copy_fields() needs to copy a non-aggregate function value,
it doesn't construct an Item_copy directly. Instead, it calls
Type_handler::create_item_copy, which is a kind of factory. The base
Type_handler::create_item_copy returns Item_copy_string. Some type
handlers override it, like timestamp and fixed binary. However, the
numeric type handlers (e.g., float, double, int, decimal) did not, so
they fell through to that base and got Item_copy_string.

A SELECT that aggregates will copy each non aggregate function value
into a temporary table through an Item_copy object, whose concrete
type is chosen by the create_item_copy method on the value's type
handler. For numeric types that method returned Item_copy_string,
which stores the value as text. A FLOAT keeps only FLT_DIG
significant digits as text, too few to reproduce its 24 bit mantissa,
so the copied value differed from the original. With one row per
group, CAST(c1 AS FLOAT) - MIN(CAST(c1 AS FLOAT)) returned a large
number instead of zero.

Add Item_copy_real with Item_copy_float and Item_copy_double variants
that keep the value as a double, the same way Item_cache_real does, and
let the float and double type handlers create them. This mirrors the
existing copy items for timestamp and fixed binary types.

947931d1c412...

947931d1c4129eb497598a2c4097e96b4c797f00

Fariha Shaikh

farihais@amazon.com

MDEV-39931 Fix main.socket_conflict failure when running as root

The test directly executes $MYSQLD via --exec, bypassing MTR's automatic
--user=root injection. In GitLab CI containers where tests run as root,
mariadbd refuses to start and the test fails.

Skip the test when running as root, matching the existing approach used
by the related main.bad_startup_options test.

All new code of the whole pull request, including one or several files
that are either new files or modified ones, are contributed under the
BSD-new license. I am contributing on behalf of my employer Amazon Web
Services, Inc.

894e2175dce8...

894e2175dce8c0e8a1314c25281d84063d9d894b

Thirunarayanan Balathandayuthapani

thiru@mariadb.com

MDEV-39963 InnoDB system tablespace autoshrink fails when the tail extent is an empty XDES_FREE_FRAG extent

Problem:
========
- The InnoDB system tablespace fails to autoshrink even when it is
almost entirely free. Defragmentation reports success but no
space is reclaimed, and the log shows the high-water mark pinned
at the end of the file.

fsp_traverse_extents(): when locating the last used extent,
descends from the end of the tablespace and treats an extent
as reclaimable only when it is XDES_FREE, or the descriptor-page
extent (XDES_FREE_FRAG with two used pages).
Every other XDES_FREE_FRAG extent stops the scan.

An XDES_FREE_FRAG extent with zero used pages can legitimately
exist on disk in tablespaces written by server versions between
commit 0b47c126e31 (MDEV-13542) and commit 7737f15f874 (MDEV-31333).
In that window fsp_free_page() evaluated xdes_get_n_used()
before clearing the freed page's XDES_FREE_BIT, so freeing the
last used page of a fragment extent left the empty extent on the
FSP_FREE_FRAG list instead of moving it to FSP_FREE.
7737f15f874 restored the correct ordering, but pre-existing
data files may still carry such extents.

Such an empty extent is logically identical to XDES_FREE, but
fsp_traverse_extents() mistook it for a used extent, pinned
last_used_extent at end-of-file, and the shrink reclaimed nothing.

Solution:
========
fsp_traverse_extents(): Treat an XDES_FREE_FRAG extent with no used
pages (n_used == 0) the same as XDES_FREE

8945d56739e6...

8945d56739e6eaacb674f145d60eb607a8b43a1a

Rucha Deodhar

rucha.deodhar@mariadb.com

MDEV-34723: NEW and OLD in a trigger as row variables

Implementation:
NEW and OLD represent the entire table row. So it can be thought of as
list of Item_trigger_field. When we are in a trigger and NEW or OLD is
encountered, create Item_trigger_row object with same constructor as
Item_trigger_field, it will also be used later while creating
Item_trigger_field objects. Populate the m_fields list while
fixing fields. Create a corresponding instruction sp_instr_set_trigger_row
which will be used to set the values

85ec122c15b7...

85ec122c15b76ded6c384e06b6d6116e51d3de8f

Kristian Nielsen

knielsen@knielsen-hq.org

Merge 10.11 to 11.4

Signed-off-by: Kristian Nielsen <[email protected]>

72e3e8e1573b...

72e3e8e1573bd5c81a1f0776f24026a31c0a3857

Yuchen Pei

ycp@mariadb.com

MDEV-39805 MDEV-39790 signal EOF in "unordered" partition index scans

This is a follow-up fix to MDEV-39535
96531691c18ebea4993454ee9d355e60ad9bea07 and MDEV-20195
8721a00dd38dc0aa1514a3b5ca8c95c6e94af1c9.

In those two fixes we added check for m_part_spec.start_part ==
NO_CURRENT_PART_ID (a typical sign of EOF) in
ha_partition::handle_unordered_prev and
ha_partition::handle_unordered_next respectively. However, because of
MDEV-37330 496aecf9251e3d74b83285f4e787daa9d8c37ae4, we skip a call to
handle_unordered_scan_next_partition inside these functions on
EOF (error == HA_ERR_END_OF_FILE and we have run out of partitions to
scan), whereas prior to MDEV-37330
496aecf9251e3d74b83285f4e787daa9d8c37ae4, it would call
handle_unordered_scan_next_partition which sets
m_part_spec.start_part= NO_CURRENT_PART_ID.

So in this patch we restore setting m_part_spec.start_part to
NO_CURRENT_PART_ID to signal EOF.

66430af54fe2...

66430af54fe2da9cce56d81f891971f4f9cbbbb1

drrtuy

drrtuy@gmail.com

chore: TPC-H automation and packaging section update in README.

6406809014db...

6406809014dba7b0efa996adc7d7bf463ea0637e

Marko Mäkelä

marko.makela@mariadb.com

MDEV-39862: Add TODO comments about MDEV-36828

This addresses a review comment
by Thirunarayanan Balathandayuthapani

5d3abcc293f2...

5d3abcc293f2a5f9cc098ba21b75475889f50808

Thirunarayanan Balathandayuthapani

thiru@mariadb.com

533e866f8620...

533e866f862050ba32930909a8be8804dba079d0

Sergei Golubchik

serg@mariadb.org

MDEV-40065 bundled command-line clients don't always enforce length limits on server results

replace strmov/strxmov with strnmov, strmake, or my_snprintf

420e28ce2846...

420e28ce2846e8a4050b1c01dc9e135206b428cc

Kristian Nielsen

knielsen@knielsen-hq.org

Merge 11.4 to 11.8

Signed-off-by: Kristian Nielsen <[email protected]>

40d62288111b...

40d62288111b511f7002561947869aa732e25f54

Kristian Nielsen

knielsen@knielsen-hq.org

Fix inconsistent terminology

Signed-off-by: Kristian Nielsen <[email protected]>

3a29e4fa8c7d...

3a29e4fa8c7d8dcd5cb5222c6b82d9d673dd9ddc

Marko Mäkelä

marko.makela@mariadb.com

MDEV-39971: generate_option_list may invoke pclose(nullptr)

call_mariadbd(): Remove the redundant and potentially harmful call
of my_pclose(f) when my_popen() failed. At least in some versions
of GNU libc the function pclose(3) is declared with
__attribute__((nonnull)).

Fixes up commit 7828fb475b00d4211ba8ce1ec6e9a1b06e0fc39b (MDEV-32745)

37964c65fa3a...

37964c65fa3aa29ff7e54f057fa24f3f60fc0547

Kristian Nielsen

knielsen@knielsen-hq.org

Merge 12.3 to 13.0

This patch to mysqldump is only merged partially:

commit 666ff1db3669a02025ae33d2cd1dfd263095c18d
Author: Monty <[email protected]>
Date: Fri May 15 16:39:35 2026 +0300

MDEV-39519 Change mariadb-dump to use SHOW REPLICA STATUS instead of SHOW SLAVE STATUS

The functionality for --dump-slave cannot be merged as the underlying
functionality was removed in this commit:

commit 21c940d5996d9a72324642a92b8bd20f71de5590
Author: ParadoxV5 <[email protected]>
Date: Thu Dec 4 21:22:20 2025 -0700

MDEV-37146 mariadb-dump: Replace SHOW SLAVE STATUS usage with information_schema.SLAVE_STATUS

Signed-off-by: Kristian Nielsen <[email protected]>

321bac944993...

321bac944993f4e940f2de2b294c5c1d013eee53

Vladislav Vaintroub

vvaintroub@gmail.com

Appveyor CI - use VS2026 image

Also, remove before_build logic trying to determine if newer commit on the
same branch exists. This takes a lot of time on outdated repos.

The same effect can be done by reducing the clone_depth to 1.
It will fail if newer commits exist, and is actually prefered - the amount
of pushes grew significantly in last time, thus fail-faster is a better
strategy to reduce build queue length.

115e5e83b7a2...

115e5e83b7a2ad788cd7f731982164594ed15ec7

Dave Gosselin

dave.gosselin@mariadb.com

MDEV-36891: std::make_pair compiler note

Construct the pair in place with emplace_back to silence the build
note.

039def0efdd9...

039def0efdd9498e62edc6667a824c60c746e0bf

Kristian Nielsen

knielsen@knielsen-hq.org

Merge 13.0 to main

Signed-off-by: Kristian Nielsen <[email protected]>

00047bff26ef...

00047bff26efd6178db4b06998e030ee5986880f

PranavKTiwari

pranav.tiwari@mariadb.com

MDEV-22943: Assertion 'marked_for_read()' failed in Field_varstring::val_str on CHECKSUM TABLE

use_all_stored_columns() cleared read_set bits for all generated
columns, including persistent generated columns.

This caused stored generated columns to be accessed without being
marked for read, leading to a marked_for_read() assertion during
CHECKSUM TABLE EXTENDED.

Only exclude non-stored virtual generated columns from read_set.