After MDEV-12353, the consistency check that I originally added for
commit 1b9fe0bbac
(InnoDB Plugin for MySQL 5.1) started randomly failing.
It turns out that the IMPORT TABLESPACE code was always incorrect:
it did not update the (redundantly stored) tablespace ID
in index tree root pages. It only does that for page headers
and BLOB pointers.
PageConverter::update_index_page(): Update the tablespace ID
in the BTR_SEG_TOP and BTR_SEG_LEAF of index root pages.
This is a backport of commit b8b3edff13.
For some reason, when mbstream is compiled with GCC 10,
ld will fail to link it due to a multiple definition of
datasink_buffer. The code appears to have been like that
since the very beginning.
Let us remove the duplicate dummy definition.
Windows GNU patch 2.7.6 is ok without it.
So account for the old buildbot version for now.
Linux works without it.
--binary fails on FreeBSD-12.0:
$ patch --version
patch 2.0-12u11 FreeBSD
$ patch --binary
patch: unrecognized option `--binary'
This reverts commit 1749a68968.
The reason why we need --binary for patch is because of a bug in
patch.exe 2.5.9. We need to supply binary otherwise the patch program
crashes.
This causes problems on FreeBSD which doesn't have a patch
that supports this.
Linux and Windows don't require it either.
Was added in c39877071a without
explaination.
Fixed a couple of race conditions in the test case to ensure stable order
of events. Also removed all sleeps. Test execution time is down from 18s
to 0.15s.
On disconnect audit event is triggered after control is returned to
mysqltest client. Which means mysqltest may issue more commands
concurrently before disconnect is actually logged.
Similar problem happens with regular query execution: an event is
triggered after control is returner to the client. Which may end
up with unstable order of events in different connections.
Delayed insert rows are enqueued separately and can either be combined
into single event or go as separate events. Reduced number of inserted
rows to 1 to stabilize result.
Also backported 2b3f6ab from 10.5.
When index_merge_sort_union is turned off only ror scans were considered for range
scans, which is wrong.
To fix the problem ensure both ror scans and non ror scans are considered for range
access
When trying to start mariadb via systemctl, WSREP failed
to start mysqld for wsrep recovery, because the binary
"galera-recovery" is neither searching the mysqld in the
same folder as the binary itself nor in the path variable
but instead expects the root to be /usr/local/mysql.
This fix changes the current directory to the desired
directory before starting mysqld.
cmake module name is BZip2, not BZIP2.
variable names are WITH_ROCKSDB_zstd and WITH_ROCKSDB_snappy.
it's SNAPPY_FOUND in older cmake (and both in newer cmake).
also, make it verbose.
was restored.
Optionally rollback prepared XA's on "mariabackup --prepare".
The fix MUST NOT be ported on 10.5+, as MDEV-742 fix solves the issue for
slaves.
A new parameter has been added called xml-report, with which the
filename of the XML file is given to which the XML result is
written. There is also xml-package for adding a package value in
the XML output. Example usage:
./mysql-test-run.pl main.events_bugs innodb.count_distinct
main.explain_json innodb.file_format_defaults json.json_no_table
--suite=main,innodb,json --force --xml-report=build123456789.xml
--xml-package=simpletestrun
It took me a long time to debug why my configs were not being loaded,
and judging from online discussions I'm not the only one. Make the
comment in the default my.cnf a bit more helpful.
The !includedir directive is implemented in mysys/my_default.c.
- f_extensions[] is a list of file extensions. It includes .ini and .cnf
on Windows, and only .cnf on all other platforms.
- search_default_file_with_ext() contains the !includedir directive. It
filters files in the directory to those matching f_extensions[].
This file should only be applicable on Unix-like platforms, so only
files with the .cnf extension are read.
Closes#1485
Let us enable all GCC and clang warnings independently of the
MYSQL_MAINTAINER_MODE setting for both Debug and RelWithDebInfo
builds, and have MYSQL_MAINTAINER_MODE only enable -Werror.
The default setting of MYSQL_MAINTAINER_MODE=AUTO will continue
to apply the -Werror only to CMAKE_BUILD_TYPE=Debug. To build
a debug version without -Werror, MYSQL_MAINTAINER_MODE=OFF can
be used.
The test main.mysqltest could crash or hang with
cmake -DWITH_ASAN=ON builds. The reason appears to be
a memory leak, which was found out by manually invoking
echo --replace_regex a > file
ASAN_OPTIONS=log_path=/dev/shm/asan mysqltest ... < file
and then examining the /dev/shm/asan.* file.
Improve the test that was imported and adapted for MariaDB in
commit fb217449dc.
row_undo_step(): Move the DEBUG_SYNC point from trx_rollback_for_mysql().
This DEBUG_SYNC point is executed after rolling back one row.
trx_rollback_for_mysql(): Clarify the comments that describe the scenario,
and remove the DEBUG_SYNC point.
If the statement "if (trx->has_logged_persistent())" and its body are
removed from trx_rollback_for_mysql(), then the test
innodb.xa_recovery_debug will fail because the transaction would still
exist in the XA PREPARE state. If we allow the XA COMMIT statement
to succeed in the test, we would observe an incorrect state of the
XA transaction where the table would contain row (1,NULL). Depending
on whether the XA transaction was committed, the table should either
be empty or contain the record (1,1). The intermediate state of
(1,NULL) should never be observed after completed recovery.
Adapt the test that was added in
mysql/mysql-server@6b65d9032c
but omitted in commit 2e814d4702.
Instead of triggering a log checkpoint, we will only trigger
a redo log flush before killing the server.
Note: the mtr.commit() call in trx_rollback_for_mysql()
will not actually make the undo log header page state change durable.
A call to log_write_up_to(mtr.commit_lsn(), true) would do that.
It is unclear what the originally reported bug scenario was.
As long as innobase_rollback_by_xid() will not return without
ensuring that the redo log has been durably written, we should be safe.
When binlog is disabled, WSREP will not behave correctly when
SAVEPOINT ROLLBACK is executed since we don't register handlers for such case.
Fixed by registering WSREP handlerton for SAVEPOINT related commands.
FOREIGN_KEY_CHECKS is disabled
- dict_foreign_find_index() can return NULL if InnoDB already dropped
the foreign index when FOREIGN_KEY_CHECKS is disabled.
If async slave thread (slave SQL handler), becomes a BF victim, it may occasionally happen that rollbacker thread is used to carry out the rollback instead of the async slave thread.
This can happen, if async slave thread has flagged "idle" state when BF thread tries to figure out how to kill the victim.
The issue was possible to test by using a galera cluster as slave for external master, and issuing high load of conflicting writes through async replication and directly against galera cluster nodes.
However, a deterministic mtr test for the "conflict window" has not yet been worked on.
The fix, in this patch makes sure that async slave thread state is never set to IDLE. This prevents the rollbacker thread to intervene.
The wsrep_query_state change was refactored to happen by dedicated function to make controlling the idle state change in one place.
redo log during recovery
- InnoDB unnecessarily reads the page even though it has fully initialized
buffered redo log records. Allow the page initialization redo log to
apply for the page in buf_page_get_gen() during recovery.
- Renamed buf_page_get_gen() to buf_page_get_low()
- Newly added buf_page_get_gen() will check for buffered redo log for
the particular page id during recovery
- Added new function buf_page_mtr_lock() which basically latches the page
for the given latch type.
- recv_recovery_create_page() is inline function which creates a page
if it has page initialization redo log records.
Update submodule, change the source file list
rename CACHE_LINE_SIZE in ut0counter, so it does not conflics with
rocksdb sources, which also defines this constant now.
Making a linked list of dtuple_t is needed only for inserting
records. It's better to store tuples in a non-intrusive
container to not affect all other use cases of dtuple_t
dtuple_t::tuple_list: removed, it was 2 * sizeof(void*) bytes
ins_node_t::entry_list: now it's std::vector<dtuple_t*>
ins_node_t::entry: now it's std::vector<dtuple_t*>::iterator
DBUG_EXECUTE_IF("row_ins_skip_sec": this dead code removed
