This bug happened when the HEAP temporary table used for the derived table
created for a derived handler of a remote engine of the federated type
became full and was converted to an Area table. For this conversion
the tmp_table_param parameter must be always taken from the select_unit
object created for the result of the derived table.
Analysis:
========
"mysqlbinlog -v" option will reconstruct row events and display them as
commented SQL statements. If this option is given twice, the output includes
comments to indicate column data types and some metadata.
`log_event_print_value` is the function reponsible for printing values and
their types. This function doesn't handle GEOMETRY type. Hence the above error
gets printed.
Fix:
===
Add support for GEOMETRY datatype.
- During insertion of clustered inde, InnoDB does the check for
foreign key constraints. It rebuild the tuple based on foreign
column names when there is no foreign index. While rebuilding,
InnoDB should ignore the dropped columns.
Before closing file handles, we really must wait until there are no
pending os_file_flush(). This was missed in
commit b1ab211dee (MDEV-15053)
where we changed the following:
fil_node_close_to_free(): Wait for n_pending==0. Because we no longer
do an extra lookup of the tablespace between fil_io() and the
completion of the operation, we must give fil_node_t::complete_io() a
chance to decrement the counter.
When maria_drop_all_indexes() reset the key files, it didn't flush
the page cache. This confused the cache as there where old key blocks
marked with LSN in it and repair tried to overwrite these with PLAIN
blocks which is not allowed.
Fixes also:
MDEV-22674 Server crash in compare_bin ... restore_table_state_after_repair
The bug was that the 'can_enable_index' variable in MyISAM and Aria was
not properly set and reset for bulk insert.
Because of this, insert...select was trying to recreate indexes while
another thread was using it, causing crashes in page cache.
The reason was that during MyISAM parallel repair two threads used the
same changed TABLE object to compute virtual columns
Fixed by adding a mutex in compute_vcols()
The problem was that the server was calling virtual functions on a record
that was not initialized with new data.
This happened when fill_record() was aborted in the middle because an
error in save_val() or save_in_field()
MDEV-17109 rpl.rpl_parallel_retry fails in buildbot with timeout
I was not able to prove that this fix works, but at least it simplifies
the problem as it removes some possible timing issues.
MDEV-21953 deadlock between BACKUP STAGE BLOCK_COMMIT and parallel
replication
Fixed by partly reverting MDEV-21953 to put back MDL_BACKUP_COMMIT locking
before log_and_order.
The original problem for MDEV-21953 was that while a thread was waiting in
for another threads to commit in 'log_and_order', it had the
MDL_BACKUP_COMMIT lock. The backup thread was waiting to get the
MDL_BACKUP_WAIT_COMMIT lock, which blocks all new MDL_BACKUP_COMMIT locks.
This causes a deadlock as the waited-for thread can never get past the
MDL_BACKUP_COMMIT lock in ha_commit_trans.
The main part of the bug fix is to release the MDL_BACKUP_COMMIT lock while
a thread is waiting for other 'previous' threads to commit. This ensures
that no transactional thread keeps MDL_BACKUP_COMMIT while waiting, which
ensures that there are no deadlocks anymore.
When commit 0fd3def284 removed
the MLOG_UNDO_ERASE_END record in MariaDB 10.3.3 in an attempt
to reduce our redo log volume, it introduced technical debt for
commit 56f6dab1d0 (MDEV-21174) and
commit 7ae21b18a6 (MDEV-12353),
which optimized mtr_t::write() (née mlog_write_ulint()) so
that the initial bytes that are unchanged are not logged.
trx_undo_report_row_operation(): Write a log record for the
memset() operation if the page is not going to be freed.
When process is aborted WSREP deinitialization and memory cleanup is not properly handled. Other reported leaks are in Galera library and should be fixed there.
Leave debian/additions/mysqlreport as #!/usr/bin/perl
Acknowledge that `env perl` is a hack, a complete fix
needs to consider which path perl is at and insert into
these scripts.
The usefulness of these scripts is questionable.
Passing a null pointer to a nonnull argument is not only undefined
behaviour, but it also grants the compiler the permission to optimize
away further checks whether the pointer is null. GCC -O2 at least
starting with version 8 may do that, potentially causing SIGSEGV.
Shifting a 16-bit type by 16 bits is undefined behaviour.
The result is at least 32 bits, so let us cast the shift operand
to a wider type before shifting.
For some reason, adding -fsanitize=undefined (cmake -DWITH_UBSAN=ON)
to the compilation flags will cause even more warnings to be emitted.
The warnings do look bogus, but the code can be simplified.
For some reason, adding -fsanitize=undefined (cmake -DWITH_UBSAN=ON)
to the compilation flags will cause even more warnings to be emitted.
The warning was a bogus one:
tests/mysql_client_test.c:8632:22: error: '%d' directive writing between
1 and 11 bytes into a region of size 9 [-Werror=format-overflow=]
8632 | sprintf(field, "c%d int", i);
| ^~
tests/mysql_client_test.c:8632:20: note: directive argument
in the range [-2147483648, 999]
The warning does not take into account that the lower bound of the
variable actually is 0. But, we can help the compiler and use an
unsigned variable.
It turns out that we must check for DISCARD TABLESPACE both
when the table is being rebuilt and when the AUTO_INCREMENT
value of the table is being added.
This was caught by the test innodb.alter_missing_tablespace.
Somehow I failed to run all tests. Sorry!
The statement ALTER TABLE...DISCARD TABLESPACE is problematic,
because its designed purpose is to break the referential integrity
of the data dictionary and make a table point to nowhere.
ha_innobase::commit_inplace_alter_table(): Check whether the
table has been discarded. (This is a bit late to check it, right
before committing the change.) Previously, we performed this check
only in a specific branch of the function commit_set_autoinc().
Note: We intentionally allow non-rebuilding ALTER TABLE even if
the tablespace has been discarded, to remain compatible with MySQL.
(See the various tests with "wl5522" in the name, such as
innodb.innodb-wl5522.)
The test case would crash starting with 10.3 only, but it does not hurt
to minimize the code and test difference between 10.2 and 10.3.
dict_load_table_low(): Copy the 'discarded' flag to file_unreadable.
This allows to avoid a potentially harmful call to dict_stats_init()
in ha_innobase::open().
