TrxInInnoDB should be constructed to track if a transaction is executing
inside InnoDB code i.e. it is like a gate between Server and InnoDB
::rnd_next() is called from Server and thus construct TrxInInnoDB
also there.
Applied suggested clean-up to TrxInInnoDB class functions enter()
and exit(). Note that exactly original did not work for enter().
The InnoDB temporary tablespace is only usable if innodb_read_only=OFF.
It is useless to create the tablespace in read-only mode, because
CREATE TEMPORARY TABLE is disallowed if innodb_read_only, and nothing
can we written to the temporary tablespace if no temporary tables
can be created.
The latest attempted fix for
MDEV-11802 InnoDB purge fails to start when there is work to do
is not complete.
For now, work around the issue by making SHOW ENGINE INNODB STATUS
call srv_wake_purge_thread_if_not_active() to wake up the purge.
rw_lock_free_func(): Assign lock->magic_n = 0 in debug builds,
because sometimes, the assignment in the explicit destructor call
is being optimized away.
This was verified when using GCC 6.3.0 with -O3 -g -mtune=native
on a Debian GNU/Linux 9.0 system on AMD64.
Use uint32_t for the encryption key_id.
When filling unsigned integer values into INFORMATION_SCHEMA tables,
use the method Field::store(longlong, bool unsigned)
instead of using Field::store(double).
Fix also some miscellanous type mismatch related to ulint (size_t).
When MySQL 5.7.9 was merged to MariaDB 10.2.2, an important
debug assertion was omitted from mlog_write_initial_log_record_low().
mlog_write_initial_log_record_low(): Put back the assertion
mtr_t::is_named_space().
fil_crypt_start_encrypting_space(), fil_crypt_rotate_page():
Call mtr_t::set_named_space() before modifying any pages.
fsp_flags_try_adjust(): Call mtr_t::set_named_space(). This additional
breakage was introduced in the merge of MDEV-11623 from 10.1. It was
not caught because of the missing debug assertion in
mlog_write_initial_log_record_low().
Remove some suppressions from the encryption.innodb-redo-badkey test.
Added a new file ha_xtradb.h where XtraDB parameters are defined. This
file is included in two places to avoid too intrusive change to
ha_innodb.cc that would make future merges harder.
innodb_show_locks_held and innodb_show_verbose_locks should be
implemented (but on different commit).
This fixes the cmake -DWITH_ASAN test failure that was mentioned
in commit f9cc391863 (merging
MDEV-12253 from 10.1 to 10.2).
fil_parse_write_crypt_data(): If the tablespace is not found,
invoke fil_space_destroy_crypt_data(&crypt_data) to properly
free the created object.
With this, the test encryption.innodb-redo-badkey still reports
"Missing MLOG_FILE_NAME or MLOG_FILE_DELETE before MLOG_CHECKPOINT"
but does not fail. The misleading message should be corrected,
maybe as part of MDEV-12699.
This bug was introduced in the fix of MDEV-12123, which invoked
page_zip_write_header() in the wrong way.
page_zip_write_header(): Assert that the length is not zero, to
be compatible with page_zip_parse_write_header().
btr_root_raise_and_insert(): Update the uncompressed page and then
invoke page_zip_write_header() with the correct length.
row_undo_mod_parse_undo_rec(): Relax the too strict assertion and
correct the comment.
innodb.innodb-blob: Force a flush of the redo log right before
killing the server, to ensure that the code path gets exercised.
(The bogus debug assertion failed on the rollback of the statement
UPDATE t3 SET c=REPEAT('j',3000) WHERE a=2 which did not modify
any indexes before the server was killed.)
Reintroduce 3d1a7cba71
that removed the maths from ut_delay.
Existing definitions of UT_COMPILE_BARRIER used.
Signed-off-by: Daniel Black <daniel.black@au.ibm.com>
ha_innobase::check_if_supported_inplace_alter(): For now, reject
ALGORITHM=INPLACE when a non-constant DEFAULT expression is specified
for ADD COLUMN or for changing a NULL column to NOT NULL.
Later, we should evaluate the non-constant column values in these cases.
The function buf_read_page_low() invokes fil_io(), which can return
DB_ERROR when the requested page is out of bounds (such as when
restoring a buffer pool dump). The callers should be handling that.
We only want to retry posix_fallocate() on EINTR as long as the system
is not being shut down. We do not want to retry on any other (hard) error.
Thanks to Jocelyn Fournier for quickly noticing the mistake in my
previous commit.
Also, include fixes by Vladislav Vaintroub to the
aws_key_management plugin. The AWS C++ SDK specifically depends on
OPENSSL_LIBRARIES, not generic SSL_LIBRARIES (such as YaSSL).
Problem was two race condtion in Aria page cache:
- find_block() didn't inform free_block() that it had released requests
- free_block() didn't handle pinned blocks, which could happen if
free_block() was called as part of flush. This is fixed by not freeing
blocks that are pinned. This is safe as when maria_close() is called
when last thread is using a table, there can be no pinned blocks. For
other flush calls it's safe to ignore pinned blocks.
During merge of innodb code QC invalidation was removed from innodb part but not added to server part.
We decided to keep it in innodb to keep server/engine interface the same.
This only merges MDEV-12253, adapting it to MDEV-12602 which is already
present in 10.2 but not yet in the 10.1 revision that is being merged.
TODO: Error handling in crash recovery needs to be improved.
If a page cannot be decrypted (or read), we should cleanly abort
the startup. If innodb_force_recovery is specified, we should
ignore the problematic page and apply redo log to other pages.
Currently, the test encryption.innodb-redo-badkey randomly fails
like this (the last messages are from cmake -DWITH_ASAN):
2017-05-05 10:19:40 140037071685504 [Note] InnoDB: Starting crash recovery from checkpoint LSN=1635994
2017-05-05 10:19:40 140037071685504 [ERROR] InnoDB: Missing MLOG_FILE_NAME or MLOG_FILE_DELETE before MLOG_CHECKPOINT for tablespace 1
2017-05-05 10:19:40 140037071685504 [ERROR] InnoDB: Plugin initialization aborted at srv0start.cc[2201] with error Data structure corruption
2017-05-05 10:19:41 140037071685504 [Note] InnoDB: Starting shutdown...
i=================================================================
==5226==ERROR: AddressSanitizer: attempting free on address which was not malloc()-ed: 0x612000018588 in thread T0
#0 0x736750 in operator delete(void*) (/mariadb/server/build/sql/mysqld+0x736750)
#1 0x1e4833f in LatchCounter::~LatchCounter() /mariadb/server/storage/innobase/include/sync0types.h:599:4
#2 0x1e480b8 in LatchMeta<LatchCounter>::~LatchMeta() /mariadb/server/storage/innobase/include/sync0types.h:786:17
#3 0x1e35509 in sync_latch_meta_destroy() /mariadb/server/storage/innobase/sync/sync0debug.cc:1622:3
#4 0x1e35314 in sync_check_close() /mariadb/server/storage/innobase/sync/sync0debug.cc:1839:2
#5 0x1dfdc18 in innodb_shutdown() /mariadb/server/storage/innobase/srv/srv0start.cc:2888:2
#6 0x197e5e6 in innobase_init(void*) /mariadb/server/storage/innobase/handler/ha_innodb.cc:4475:3
(from: http://buildbot.askmonty.org/buildbot/builders/p8-rhel6-bintar/builds/820/steps/test/logs/stdio)
Errors like the following indicate a potential endian storage issue:
rocksdb.rocksdb_range w1 [ fail ]
Test ended at 2017-04-27 18:56:11
CURRENT_TEST: rocksdb.rocksdb_range
--- /home/buildbot/maria-slave/p8-rhel6-bintar/build/storage/rocksdb/mysql-test/rocksdb/r/rocksdb_range.result 2017-04-27 17:41:27.740050347 -0400
+++ /home/buildbot/maria-slave/p8-rhel6-bintar/build/storage/rocksdb/mysql-test/rocksdb/r/rocksdb_range.reject 2017-04-27 18:56:11.230050346 -0400
@@ -25,15 +25,15 @@
select * from t2 force index (a) where a=0;
pk a b
0 0 0
-1 0 1
-2 0 2
-3 0 3
-4 0 4
-5 0 5
-6 0 6
-7 0 7
-8 0 8
-9 0 9
+16777216 0 1
+33554432 0 2
+50331648 0 3
+67108864 0 4
+83886080 0 5
+100663296 0 6
+117440512 0 7
+134217728 0 8
+150994944 0 9
# The rest are for code coverage:
explain
select * from t2 force index (a) where a=2;
@@ -41,23 +41,23 @@
1 SIMPLE t2 ref a a 4 const #
select * from t2 force index (a) where a=2;
pk a b
-20 2 20
-21 2 21
-22 2 22
-23 2 23
-24 2 24
-25 2 25
-26 2 26
-27 2 27
-28 2 28
-29 2 29
+335544320 2 20
+352321536 2 21
+369098752 2 22
+385875968 2 23
+402653184 2 24
+419430400 2 25
+436207616 2 26
+452984832 2 27
+469762048 2 28
+486539264 2 29
explain
select * from t2 force index (a) where a=3 and pk=33;
id select_type table type possible_keys key key_len ref rows Extra
1 SIMPLE t2 const a a 8 const,const #
select * from t2 force index (a) where a=3 and pk=33;
pk a b
-33 3 33
+553648128 3 33
select * from t2 force index (a) where a=99 and pk=99;
pk a b
select * from t2 force index (a) where a=0 and pk=0;
...
Signed-off-by: Daniel Black <daniel.black@au.ibm.com>
This regression was caused by MDEV-12467 encryption.create_or_replace
hangs during DROP TABLE, where if table->is_stopping() (i.e. when
tablespace is dropped) background key rotation thread calls
fil_crypt_complete_rotate_space to release space and stop rotation.
However, that function does not decrease number of rotating
threads if table->is_stopping() is true.
- Update rocksdb submodule to revision
d616ebea23fa88cb9c2c8588533526a566d9cfab
- Normally this should be done by doing a merge from upstream
MyRocks, but now we are just updating rocksdb, MyRocks merge will
follow later.
- Apply a part of 4f6f072f6c74513087004755508eb6d6c432c5c1
use_direct_writes was renamed to use_direct_io_for_flush_and_compaction
- Update build_rocksdb.cmake - RocksDB have moved files around
innochecksum uses global variables. great, let's use them all the
way down, instead of passing them as arguments to innodb internals,
conditionally modifying function prototypes with #ifdefs
|| kv >= crypt_data->min_key_version,
encryption.innodb_encryption_tables failed in buildbot.
Now that key_version is not stored when page is read to
buf_page_t::key_version but always read from actual page
this assertion is not always valid.
This fixes a regression caused by MDEV-12428.
When we introduced a variant of fil_space_acquire() that could
increment space->n_pending_ops after space->stop_new_ops was set,
the logic of fil_check_pending_operations() was broken.
fil_space_t::n_pending_ios: A new field to track read or write
access from the buffer pool routines immediately before a block
write or after a block read in the file system.
fil_space_acquire_for_io(), fil_space_release_for_io(): Similar
to fil_space_acquire_silent() and fil_space_release(), but
modify fil_space_t::n_pending_ios instead of fil_space_t::n_pending_ops.
Adjust a number of places accordingly, and remove some redundant
tablespace lookups.
The following parts of this fix differ from the 10.2 version of this fix:
buf_page_get_corrupt(): Add a tablespace parameter.
In 10.2, we already had a two-phase process of freeing fil_space objects
(first, fil_space_detach(), then release fil_system->mutex, and finally
free the fil_space and fil_node objects).
fil_space_free_and_mutex_exit(): Renamed from fil_space_free().
Detach the tablespace from the fil_system cache, release the
fil_system->mutex, and then wait for space->n_pending_ios to reach 0,
to avoid accessing freed data in a concurrent thread.
During the wait, future calls to fil_space_acquire_for_io() will
not find this tablespace, and the count can only be decremented to 0,
at which point it is safe to free the objects.
fil_node_free_part1(), fil_node_free_part2(): Refactored from
fil_node_free().
This fixes a regression caused by MDEV-12428.
When we introduced a variant of fil_space_acquire() that could
increment space->n_pending_ops after space->stop_new_ops was set,
the logic of fil_check_pending_operations() was broken.
fil_space_t::n_pending_ios: A new field to track read or write
access from the buffer pool routines immediately before a block
write or after a block read in the file system.
fil_space_acquire_for_io(), fil_space_release_for_io(): Similar
to fil_space_acquire_silent() and fil_space_release(), but
modify fil_space_t::n_pending_ios instead of fil_space_t::n_pending_ops.
fil_space_free_low(): Wait for space->n_pending_ios to reach 0,
to avoid accessing freed data in a concurrent thread. Future
calls to fil_space_acquire_for_io() will not find this tablespace,
because it will already have been detached from fil_system.
Adjust a number of places accordingly, and remove some redundant
tablespace lookups.
FIXME: buf_page_check_corrupt() should take a tablespace from
fil_space_acquire_for_io() as a parameter. This will be done
in the 10.1 version of this patch and merged from there.
That depends on MDEV-12253, which has not been merged from 10.1 yet.
These changes are comparable to Percona's modifications in innodb in the
Percona Xtrabackup repository.
- If functions are used in backup as well as in innodb, make them non-static.
- Define IS_XTRABACKUP() macro for special handling of innodb running
inside backup.
- Extend some functions for backup.
fil_space_for_table_exists_in_mem() gets additional parameter
'remove_from_data_dict_if_does_not_exist', for partial backups
fil_load_single_table_tablespaces() gets an optional parameter predicate
which tells whether to load tablespace based on database or table name,
also for partial backups.
srv_undo_tablespaces_init() gets an optional parameter 'backup_mode'
- Allow single redo log file (for backup "prepare")
- Do not read doublewrite buffer pages in backup, they are outdated
- Add function fil_remove_invalid_table_from_data_dict(), to remove non-existing
tables from data dictionary in case of partial backups.
- On Windows, fix file share modes when opening tablespaces,
to allow mariabackup to read tablespaces while server is online.
- Avoid access to THDVARs in backup, because innodb plugin is not loaded,
and THDVAR would crash in this case.
- Do not throw output of exec command, if disable_result_log is set
save and dump it if exec fails. Need tha to meaningfully analyze
errors from mariabackup.
- rmdir now removes the entire tree. need that because xtrabackup tests
clean the whole directory.
- all filesystem modifying commands now require the argument to
be under MYSQLTEST_VARDIR or MYSQL_TMP_DIR.
