- developers - lists.mariadb.org

Re: [Maria-developers] [Commits] 203a658: MDEV-11084 Select statement with partition selection against MyISAM table opens all partitions.
by Sergei Golubchik 19 Dec '17

19 Dec '17

Hi, Alexey! On Dec 18, Alexey Botchkov wrote: > revision-id: 203a6585d3deedb36e30d5ebf388de44b6810403 (mariadb-10.1.23-51-g203a658) > parent(s): e4d10e09cf318aad237143254c45458d16009f70 > committer: Alexey Botchkov > timestamp: 2017-12-18 19:31:26 +0400 > message: > > MDEV-11084 Select statement with partition selection against MyISAM table opens all partitions. > > Now SELECT FROM t PARTITION(x) only opens the 'x' file. > The table->partition_names parameter is sent to ha_open > so it only opens the required partitions. > If the table is reopened, the > change_partition_to_open(partition_names) is called that > closes and opens partitions specified for the query. Looks ok to me, thanks! Please add some tests, can status variables show the difference in the number of opened partitions? And make sure that the whole tree compiles without partitioning. Then ok to push into 10.3 Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

1 0

Re: [Maria-developers] [Commits] 7a7ad82: MDEV-13478 Full SST sync fails because of the error in the cleaning part
by Sergei Golubchik 19 Dec '17

19 Dec '17

Hi, Sachin! Unfortunately, your comment is rather difficult to understand. What about this one: Problem: The command was: find $paths -mindepth 1 -regex $cpat -prune -o -exec rm -rf {} \+ Which was supposed to work as * skipping $paths directories themselves (-mindepth 1) * see if the dir/file name matches $cpat (-regex) * if yes - don't dive into the directory, skip it (-prune) * otherwise (-o) * remove it and everything inside (-exec) Now -exec ... \+ works like this: every new found path is appended to the end of the command line. when accumulated command line length reaches `getconf ARG_MAX` (~2Gb) it's executed, and find continues, appending to a new command line. What happens here, find appends some directory to the command line, then dives into it, and starts appending files from that directory. At some point command line overflows, rm -rf gets executed and removes the whole directory. Now find tries to continue scanning the directory that was already removed. Fix: don't dive into directories that will be recursively removed anyway, use -prune for them. Basically, we should be pruning both paths that have matched $cpat and paths that have not matched it. This is achived by pruning unconditionally, before the regex is tested: find $paths -mindepth 1 -prune -regex $cpat -o -exec rm -rf {} \+ On Dec 19, sachin wrote: > revision-id: 7a7ad82029a6c78d31d6736a562c12d02c4d968c (mariadb-galera-5.5.58-3-g7a7ad82) > parent(s): e6e026ae51a77969749de201d491a176483bbc69 > author: Sachin Setiya > committer: Sachin Setiya > timestamp: 2017-12-19 22:30:43 +0530 > message: > > MDEV-13478 Full SST sync fails because of the error in the cleaning part > > Problem:- The problem is in wsrep_sst_xtrabackup-v2.sh we use > find $ib_home_dir $ib_log_dir $ib_undo_dir $DATA -mindepth 1 -regex $cpat -prune -o -exec rm -rfv {} 1>&2 \+ > the problem is that since we have '\+' in end that means all output > will be expanded after rm -rfv . If we have really large database( > quite a no of tables with big names) then this create a problem. > This will result in calling 'rm -rvf' two time(or may be more). So non > deterministicly this might that upto directory name went to rm and > remaining was truncated. For example consider a folder xyz with lots > of files. We executed > find xyz -exec rm -rfv {} \+ > Since we have like millions of file in rm, So it will be greater then > ARG_MAX , so there will be multiple rm invocation. Say in nth invocation > this might happen rm ....... xyz/ {So it get truncated at xyz/} > Above will remove the whole xyz directory and remaining invocation > of rm will return error. How ever this type of error is non deterministic > so is bug. > Solution:- > In above example if instead of removing each file in xyz if we remove > xyz then we have our solution :). > Actually if we shift the -prune term in find we can get to solution. > Why currently find is working like this > find ( -regex && -prune) || -exec > that means if -regex is true (-prune will always return true) then exec > (and hence rm )wont work. But if regex fails then exec is applied with > out -prune which makes rm delete each single file instead of folder. > So we change the position of regex and prune then prune will always be > applied whether regex is true or not. > find (-prune && -regex) || -exec > > Patch Credit:- Serg > > --- > scripts/wsrep_sst_xtrabackup-v2.sh | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/scripts/wsrep_sst_xtrabackup-v2.sh b/scripts/wsrep_sst_xtrabackup-v2.sh > index 327c92f..00d8fe2 100644 > --- a/scripts/wsrep_sst_xtrabackup-v2.sh > +++ b/scripts/wsrep_sst_xtrabackup-v2.sh > @@ -863,7 +863,7 @@ then > > > wsrep_log_info "Cleaning the existing datadir and innodb-data/log directories" > - find $ib_home_dir $ib_log_dir $ib_undo_dir $DATA -mindepth 1 -regex $cpat -prune -o -exec rm -rfv {} 1>&2 \+ > + find $ib_home_dir $ib_log_dir $ib_undo_dir $DATA -mindepth 1 -prune -regex $cpat -o -exec rm -rfv {} 1>&2 \+ > > tempdir=$(parse_cnf mysqld log-bin "") > if [[ -n ${tempdir:-} ]];then Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

2 1

Re: [Maria-developers] 665402a: MDEV-14265 - RPMLint warning: shared-lib-calls-exit
by Sergei Golubchik 18 Dec '17

18 Dec '17

Hi, Sergey! On Dec 18, Sergey Vojtovich wrote: > revision-id: 665402aa9b83565c1cef3caa06cd998d3808f0d7 (mariadb-10.1.29-31-g665402a) > parent(s): ef9e78c9d41a5ac644f08068e3dabad948b0e30a > committer: Sergey Vojtovich > timestamp: 2017-12-18 20:23:36 +0400 > message: > > MDEV-14265 - RPMLint warning: shared-lib-calls-exit > > find_type_with_warning() client helper did exit(1) on error, exit(1) moved to > clients. It was find_type_or_exit(), and you could've had less changes if you'd replaced it with an inline function. Basically, copied the function into the header and made it static inline. > mysql_read_default_options() did exit(1) on error, error is passed through and > handled now. ok > my_str_malloc_default() did exit(1) on error, replaced my_str_ allocator > functions with normal my_malloc()/my_realloc()/my_free(). I think you need to replace my_str_malloc_default() with my_malloc(..., MYF(MY_FAE)), not my_malloc(..., MYF(0)) > sql_connect.cc did many exit(1) on hash initialisation failure. Removed error > check since my_hash_init() never fails. > > my_malloc() did exit(1) on error. Replaced with abort(). > > my_load_defaults() did exit(1) on error, replaced with abort(). I'd say, return 2 instead of exit/abort. my_load_defaults already does it, because my_search_option_files returns 2 for OOM. > my_load_defaults() still does exit(0) when invoked with --print-defaults. dunno, return -1, for example, and exit in the caller? Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

2 1

[Maria-developers] MDEV-14031 Password policy causes replication failure
by Alexander Barkov 18 Dec '17

18 Dec '17

Hi Sergei, Please review a patch for MDEV-14031. Thanks!

1 0

Re: [Maria-developers] [Commits] 2603dc0: MDEV-11084 Select statement with partition selection against MyISAM table opens all partitions.
by Sergei Golubchik 15 Dec '17

15 Dec '17

Hi, Alexey! That looked reasonable enough. I didn't find anything big. Small comments - see below. On Jun 01, Alexey Botchkov wrote: > revision-id: 2603dc0d56ec90f308678a3c002f1b75af69f554 (mariadb-10.1.23-51-g2603dc0) > parent(s): e4d10e09cf318aad237143254c45458d16009f70 > committer: Alexey Botchkov > timestamp: 2017-06-01 17:33:34 +0400 > message: > > MDEV-11084 Select statement with partition selection against MyISAM table opens all partitions. > > Now SELECT FROM t PARTITION(x) only opens the 'x' file. > The table->partition_names parameter is sent to ha_open > so it only opens the required partitions. > If the table is reopened, the > change_partition_to_open(partition_names) is called that > closes and opens partitions specified for the query. > > diff --git a/sql/ha_partition.cc b/sql/ha_partition.cc > index 29054c7..53a128c 100644 > --- a/sql/ha_partition.cc > +++ b/sql/ha_partition.cc > @@ -3481,12 +3498,16 @@ int ha_partition::open(const char *name, int mode, uint test_if_locked) > } > name_buffer_ptr+= strlen(name_buffer_ptr) + 1; > } > + file_sample= m_file[0]; are you sure that m_file[0] is opened? > } > else > { > file= m_file; > do > { > + if (!bitmap_is_set(&m_opened_partitions, file - m_file)) > + continue; > + > create_partition_name(name_buff, name, name_buffer_ptr, NORMAL_PART_NAME, > FALSE); > table->s->connect_string = m_connect_string[(uint)(file-m_file)]; > @@ -3494,20 +3515,21 @@ int ha_partition::open(const char *name, int mode, uint test_if_locked) > test_if_locked | HA_OPEN_NO_PSI_CALL))) > goto err_handler; > bzero(&table->s->connect_string, sizeof(LEX_STRING)); > - if (m_file == file) > - m_num_locks= (*file)->lock_count(); > + if (!file_sample) > + m_num_locks= (file_sample= (*file))->lock_count(); > DBUG_ASSERT(m_num_locks == (*file)->lock_count()); > - name_buffer_ptr+= strlen(name_buffer_ptr) + 1; > - } while (*(++file)); > + } while (name_buffer_ptr+= strlen(name_buffer_ptr) + 1, *(++file)); why? > } > > file= m_file; > - ref_length= (*file)->ref_length; > - check_table_flags= (((*file)->ha_table_flags() & > + ref_length= file_sample->ref_length; > + check_table_flags= ((file_sample->ha_table_flags() & > ~(PARTITION_DISABLED_TABLE_FLAGS)) | > (PARTITION_ENABLED_TABLE_FLAGS)); > while (*(++file)) > { > + if (!bitmap_is_set(&m_opened_partitions, file - m_file)) > + continue; > /* MyISAM can have smaller ref_length for partitions with MAX_ROWS set */ > set_if_bigger(ref_length, ((*file)->ref_length)); > /* > @@ -6782,6 +6809,63 @@ void ha_partition::get_dynamic_partition_info(PARTITION_STATS *stat_info, > } > > > +void ha_partition::set_partitions_to_open(List<String> *partition_names) > +{ > + m_partitions_to_open= partition_names; > +} > + > + > +int ha_partition::change_partitions_to_open(List<String> *partition_names) > +{ > + char name_buff[FN_REFLEN]; > + handler **file; > + char *name_buffer_ptr; > + int error= 0; > + > + if (m_is_clone_of) > + return 0; > + > + m_partitions_to_open= partition_names; > + if ((error= m_part_info->set_partition_bitmaps(partition_names))) > + goto err_handler; > + > + if (bitmap_cmp(&m_opened_partitions, &m_part_info->read_partitions) != 0) > + return 0; > + > + if ((error= read_par_file(table->s->normalized_path.str))) > + goto err_handler; > + > + name_buffer_ptr= m_name_buffer_ptr; > + file= m_file; > + do > + { > + int n_file= file-m_file; > + int is_open= bitmap_is_set(&m_opened_partitions, n_file); > + int should_be_open= bitmap_is_set(&m_part_info->read_partitions, n_file); > + > + if (is_open && !should_be_open) > + (*file)->ha_close(); > + else if (!is_open && should_be_open) > + { > + create_partition_name(name_buff, table->s->normalized_path.str, > + name_buffer_ptr, NORMAL_PART_NAME, FALSE); > + table->s->connect_string = m_connect_string[(uint)(file-m_file)]; > + if ((error= (*file)->ha_open(table, name_buff, m_mode, > + m_open_test_lock | HA_OPEN_NO_PSI_CALL))) > + goto err_handler; > + bzero(&table->s->connect_string, sizeof(LEX_STRING)); > + } > + } while (name_buffer_ptr+= strlen(name_buffer_ptr) + 1, *(++file)); Could you reuse that, instead of copying from ::open? Perhaps ::open() can reset the bitmap to be empty and then call ::change_partitions_to_open() ? Or the common code could be moved into a third function that these both will use... > + > + bitmap_copy(&m_opened_partitions, &m_part_info->read_partitions); > + > + clear_handler_file(); > + > +err_handler: > + return error; > +} > + > + > /** > General function to prepare handler for certain behavior. > > diff --git a/sql/partition_info.cc b/sql/partition_info.cc > index bc0db9e..7fb8f70 100644 > --- a/sql/partition_info.cc > +++ b/sql/partition_info.cc > @@ -283,6 +281,27 @@ bool partition_info::set_partition_bitmaps(TABLE_LIST *table_list) > > > /** > + Set read/lock_partitions bitmap over non pruned partitions > + > + @param table_list Possible TABLE_LIST which can contain > + list of partition names to query > + > + @return Operation status > + @retval FALSE OK > + @retval TRUE Failed to allocate memory for bitmap or list of partitions > + did not match > + > + @note OK to call multiple times without the need for free_bitmaps. > +*/ > +bool partition_info::set_partition_bitmaps_from_table(TABLE_LIST *table_list) > +{ > + List<String> *partition_names= table_list ? > + NULL : table_list->partition_names; > + return set_partition_bitmaps(partition_names); > +} This looks unused. > + > + > +/** > Checks if possible to do prune partitions on insert. > > @param thd Thread context > diff --git a/sql/sql_base.cc b/sql/sql_base.cc > index 9c56b7d..2f9ec3e 100644 > --- a/sql/sql_base.cc > +++ b/sql/sql_base.cc > @@ -2533,6 +2533,12 @@ bool open_table(THD *thd, TABLE_LIST *table_list, Open_table_context *ot_ctx) > { > DBUG_ASSERT(table->file != NULL); > MYSQL_REBIND_TABLE(table->file); > + if (table->part_info) > + { > + /* Set all [named] partitions as used. */ > + if (table->file->change_partitions_to_open(table_list->partition_names)) > + DBUG_RETURN(true); > + } > } > else > { > @@ -2549,7 +2555,8 @@ bool open_table(THD *thd, TABLE_LIST *table_list, Open_table_context *ot_ctx) > HA_TRY_READ_ONLY), > (READ_KEYINFO | COMPUTE_TYPES | > EXTRA_RECORD), > - thd->open_options, table, FALSE); > + thd->open_options, table, FALSE, > + table_list->partition_names); Please, make sure you code compiles without partitioning too. (in this case I wouldn't make the last argument of open_table_from_share conditional, but I'd rather make table_list->partition_names unconditional). > > if (error) > { > @@ -2598,9 +2605,11 @@ bool open_table(THD *thd, TABLE_LIST *table_list, Open_table_context *ot_ctx) > #ifdef WITH_PARTITION_STORAGE_ENGINE > if (table->part_info) > { > +#ifdef CODE_TO_REMOVE > /* Set all [named] partitions as used. */ > if (table->part_info->set_partition_bitmaps(table_list)) > DBUG_RETURN(true); > +#endif /*CODE_TO_REMOVE*/ why not to remove at once? > } > else if (table_list->partition_names) > { Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

1 0

Re: [Maria-developers] 14acb68: MDEV-14593 human-readable XA RECOVER.
by Sergei Golubchik 14 Dec '17

14 Dec '17

Hi, Alexey! On Dec 12, Alexey Botchkov wrote: > revision-id: 14acb687199f0d30444f8c0849c0d1c1284f3e2b (mariadb-10.2.11-26-g14acb68) > parent(s): 4d016e6ed2d32298977a66e125cbfcae39e23847 > committer: Alexey Botchkov > timestamp: 2017-12-12 14:31:41 +0400 > message: > > MDEV-14593 human-readable XA RECOVER. > > The 'data' field in the XA RECOVER resultset changed > to be charset_bin. It seems to me right and also > --binary-as-hex starts working. > The XA RECOVER FORMAT='SQL' option implemented. > It returns the XID string that fits to be an argument for the > XA ... statements. Looks ok, just a couple of comments: > +xa recover format='SQL'; Please add a test for format=SQL (without quotes). We normally document EXPLAIN FORMAT=JSON without quotes (although it works with quotes too), so XA RECOVER should work the same way. > +ER_UNKNOWN_XID_FORMAT > + eng "Unknown XID format name: '%s'" Don't add a new error message, reuse ER_UNKNOWN_EXPLAIN_FORMAT: - eng "Unknown EXPLAIN format name: '%s'" - rus "Неизвестное имя формата команды EXPLAIN: '%s'" + eng "Unknown %s format name: '%s'" + rus "Неизвестное имя формата команды %s: '%s'" note, you'll need either to remove/rephrase the russian error message, or use "XA RECOVER" as the first argument, not "XID". Ok to push (into bb-10.2-ext) with these changes. Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org

1 0

Re: [Maria-developers] [Commits] 8048670: MDEV-10715 Galera: Replicate MariaDB GTID to other nodes in the cluster
by Jan Lindström 12 Dec '17

12 Dec '17

On Mon, Dec 11, 2017 at 9:40 AM, sachin <sachin.setiya(a)mariadb.com> wrote: > revision-id: 8048670e25e5cda915ad7ce16dbd66afa6227aa5 > (mariadb-10.1.29-24-g8048670) > parent(s): de76cbdcb0728b80c0f7b453b24b0b81f653e450 > author: Sachin Setiya > committer: Sachin Setiya > timestamp: 2017-12-11 13:05:08 +0530 > message: > > MDEV-10715 Galera: Replicate MariaDB GTID to other nodes in the cluster > > Problem:- Gtid are not transferred in Galera Cluster. > > Solution:- We need to transfer gtid in the case on either when cluster is > slave/master in async replication. In normal Gtid replication gtid are > generated on > recieving node itself and it is always on sync with other nodes. Because > galera keeps > node in sync , So all nodes get same no of event groups. So the issue > arises when > say galera is slave in async replication. > A > | (Async replication) > D <-> E <-> F {Galera replication} > So what should happen is that all node should apply the master gtid but > this does > node happen, becuase node E, F does not recieve gtid from D in write set , > So what E(or F) > does is that it applies wsrep_gtid_domain_id, D server-id , E gtid next > seq no. This > generated gtid does not always work when say A has different domain id. > > So In this commit, on galera node when we see that this event is recieved > from master > we simply write Gtid_Log_Event in write_set and send it to other nodes. > > --- > mysql-test/suite/galera/r/galera_gtid_slave.result | 26 +++ > .../r/galera_gtid_slave_sst_mysqldump.result | 133 +++++++++++++ > .../galera/r/galera_gtid_slave_sst_rsync.result | 130 +++++++++++++ > mysql-test/suite/galera/t/galera_gtid_slave.cnf | 18 ++ > mysql-test/suite/galera/t/galera_gtid_slave.test | 78 ++++++++ > .../galera/t/galera_gtid_slave_sst_mysqldump.cnf | 19 ++ > .../galera/t/galera_gtid_slave_sst_mysqldump.test | 208 > +++++++++++++++++++++ > .../suite/galera/t/galera_gtid_slave_sst_rsync.cnf | 18 ++ > .../galera/t/galera_gtid_slave_sst_rsync.test | 207 > ++++++++++++++++++++ > mysql-test/suite/galera_3nodes/galera_2x3nodes.cnf | 122 ++++++++++++ > .../galera_3nodes/r/galera_gtid_2_cluster.result | 117 ++++++++++++ > .../galera_3nodes/t/galera_gtid_2_cluster.cnf | 28 +++ > .../galera_3nodes/t/galera_gtid_2_cluster.test | 143 ++++++++++++++ > sql/log.cc | 51 ++++- > sql/wsrep_mysqld.cc | 10 + > 15 files changed, 1306 insertions(+), 2 deletions(-) > > diff --git a/mysql-test/suite/galera/r/galera_gtid_slave.result > b/mysql-test/suite/galera/r/galera_gtid_slave.result > new file mode 100644 > index 0000000..40f3f1c > --- /dev/null > +++ b/mysql-test/suite/galera/r/galera_gtid_slave.result > @@ -0,0 +1,26 @@ > +START SLAVE; > +CREATE TABLE t1 (f1 INTEGER PRIMARY KEY) ENGINE=InnoDB; > +CREATE TABLE t2 (f1 INTEGER PRIMARY KEY) ENGINE=InnoDB; > +INSERT INTO t1 VALUES(1); > +begin; > +insert into t2 values(21); > +insert into t2 values(22); > +commit; > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4 > +INSERT INTO t1 VALUES(2); > +INSERT INTO t1 VALUES(3); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4,2-2-2 > +INSERT INTO t1 VALUES(4); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4,2-2-2,2-3-3 > +DROP TABLE t1,t2; > +reset master; > +STOP SLAVE; > +RESET SLAVE ALL; > +reset master; > +reset master; > diff --git a/mysql-test/suite/galera/r/galera_gtid_slave_sst_mysqldump.result > b/mysql-test/suite/galera/r/galera_gtid_slave_sst_mysqldump.result > new file mode 100644 > index 0000000..78a0a17 > --- /dev/null > +++ b/mysql-test/suite/galera/r/galera_gtid_slave_sst_mysqldump.result > @@ -0,0 +1,133 @@ > +#Connection 2 > +START SLAVE; > +include/wait_for_slave_to_start.inc > +#Connection 1 > +CREATE TABLE t2 (f1 INTEGER PRIMARY KEY, f2 int unique) ENGINE=InnoDB; > +INSERT INTO t2 VALUES(1,11); > +INSERT INTO t2 VALUES(2,22); > +INSERT INTO t2 VALUES(3,33); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4 > +include/save_master_gtid.inc > +#Connection 2 > +include/sync_with_master_gtid.inc > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4 > +INSERT INTO t2 VALUES(4,44); > +INSERT INTO t2 VALUES(5,55); > +INSERT INTO t2 VALUES(6,66); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4,2-2-3 > +#Connection 3 > +INSERT INTO t2 VALUES(7,77); > +INSERT INTO t2 VALUES(8,88); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4,2-2-3,2-3-5 > +#Connection 1 > +CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_before'); > +INSERT INTO t1 VALUES ('node1_committed_before'); > +COMMIT; > +include/save_master_gtid.inc > +#Connection 2 > +include/sync_with_master_gtid.inc > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node2_committed_before'); > +INSERT INTO t1 VALUES ('node2_committed_before'); > +COMMIT; > +#Connection 3 > +Shutting down server ... > +#Connection 2 > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_during'); > +INSERT INTO t1 VALUES ('node1_committed_during'); > +COMMIT; > +#Connection 3 > +Starting server ... > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node3_committed_after'); > +INSERT INTO t1 VALUES ('node3_committed_after'); > +COMMIT; > +#Connection 2 > +Select * from t1 order by f1; > +f1 > +node1_committed_before > +node1_committed_before > +node1_committed_during > +node1_committed_during > +node2_committed_before > +node2_committed_before > +node3_committed_after > +node3_committed_after > +#Connection 3 > +Select * from t1 order by f1; > +f1 > +node1_committed_before > +node1_committed_before > +node1_committed_during > +node1_committed_during > +node2_committed_before > +node2_committed_before > +node3_committed_after > +node3_committed_after > +#Connection 2 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-6,2-2-7,2-3-8 > +#Connection 3 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-6,2-2-7,2-3-8 > +#Connection 1 > +SET AUTOCOMMIT=ON; > +#Connection 2 > +SET AUTOCOMMIT=ON; > +#Connection 3 > +SET AUTOCOMMIT=ON; > +#Connection 2 > +STOP slave; > +include/wait_for_slave_to_stop.inc > +INSERT INTO t1 VALUES ('node2_slave_stoped'); > +#Connection 1 > +INSERT INTO t1 VALUES ('node1_normal_entry'); > +include/save_master_gtid.inc > +#Connection 2 > +INSERT INTO t1 VALUES ('node2_slave_stoped_inserted'); > +start slave; > +include/wait_for_slave_to_start.inc > +include/sync_with_master_gtid.inc > +INSERT INTO t1 VALUES ('node2_slave_started'); > +SELECT count(*) from t1; > +count(*) > +12 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-7,2-3-8,2-2-11 > +#Connection 3 > +SELECT count(*) from t1; > +count(*) > +12 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-7,2-3-8,2-2-11 > +#Connection 1 > +DROP TABLE t2,t1; > +#Connection 2 > +#Connection 3 > +#Connection 2 > +STOP SLAVE; > +RESET SLAVE; > +reset master; > +#Connection 3 > +reset master; > +#Connection 1 > +reset master; > diff --git a/mysql-test/suite/galera/r/galera_gtid_slave_sst_rsync.result > b/mysql-test/suite/galera/r/galera_gtid_slave_sst_rsync.result > new file mode 100644 > index 0000000..81fae57 > --- /dev/null > +++ b/mysql-test/suite/galera/r/galera_gtid_slave_sst_rsync.result > @@ -0,0 +1,130 @@ > +#Connection 2 > +START SLAVE; > +#Connection 1 > +CREATE TABLE t2 (f1 INTEGER PRIMARY KEY, f2 int unique) ENGINE=InnoDB; > +INSERT INTO t2 VALUES(1,11); > +INSERT INTO t2 VALUES(2,22); > +INSERT INTO t2 VALUES(3,33); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4 > +include/save_master_gtid.inc > +#Connection 2 > +include/sync_with_master_gtid.inc > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4 > +INSERT INTO t2 VALUES(4,44); > +INSERT INTO t2 VALUES(5,55); > +INSERT INTO t2 VALUES(6,66); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4,2-2-3 > +#Connection 3 > +INSERT INTO t2 VALUES(7,77); > +INSERT INTO t2 VALUES(8,88); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4,2-2-3,2-3-5 > +#Connection 1 > +CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_before'); > +INSERT INTO t1 VALUES ('node1_committed_before'); > +COMMIT; > +include/save_master_gtid.inc > +#Connection 2 > +include/sync_with_master_gtid.inc > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node2_committed_before'); > +INSERT INTO t1 VALUES ('node2_committed_before'); > +COMMIT; > +#Connection 3 > +Shutting down server ... > +#Connection 2 > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_during'); > +INSERT INTO t1 VALUES ('node1_committed_during'); > +COMMIT; > +#Connection 3 > +Starting server ... > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node3_committed_after'); > +INSERT INTO t1 VALUES ('node3_committed_after'); > +COMMIT; > +#Connection 2 > +Select * from t1 order by f1; > +f1 > +node1_committed_before > +node1_committed_before > +node1_committed_during > +node1_committed_during > +node2_committed_before > +node2_committed_before > +node3_committed_after > +node3_committed_after > +#Connection 3 > +Select * from t1 order by f1; > +f1 > +node1_committed_before > +node1_committed_before > +node1_committed_during > +node1_committed_during > +node2_committed_before > +node2_committed_before > +node3_committed_after > +node3_committed_after > +#Connection 2 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-6,2-2-7,2-3-8 > +#Connection 3 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-6,2-2-7,2-3-8 > +#Connection 1 > +SET AUTOCOMMIT=ON; > +#Connection 2 > +SET AUTOCOMMIT=ON; > +#Connection 3 > +SET AUTOCOMMIT=ON; > +#Connection 2 > +STOP slave; > +INSERT INTO t1 VALUES ('node2_slave_stoped'); > +#Connection 1 > +INSERT INTO t1 VALUES ('node1_normal_entry'); > +include/save_master_gtid.inc > +#Connection 2 > +INSERT INTO t1 VALUES ('node2_slave_stoped_inserted'); > +start slave; > +include/sync_with_master_gtid.inc > +INSERT INTO t1 VALUES ('node2_slave_started'); > +SELECT count(*) from t1; > +count(*) > +12 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-7,2-3-8,2-2-11 > +#Connection 3 > +SELECT count(*) from t1; > +count(*) > +12 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-7,2-3-8,2-2-11 > +#Connection 1 > +DROP TABLE t2,t1; > +#Connection 2 > +#Connection 3 > +#Connection 2 > +STOP SLAVE; > +RESET SLAVE ALL; > +reset master; > +#Connection 3 > +reset master; > +#Connection 1 > +reset master; > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave.cnf > b/mysql-test/suite/galera/t/galera_gtid_slave.cnf > new file mode 100644 > index 0000000..409d0d1 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave.cnf > @@ -0,0 +1,18 @@ > +!include ../galera_2nodes_as_slave.cnf > + > +[mysqld] > +log-bin=mysqld-bin > +log-slave-updates > +binlog-format=ROW > + > +[mysqld.1] > +gtid-domain-id=1 > +[mysqld.2] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > +[mysqld.3] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > + > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave.test > b/mysql-test/suite/galera/t/galera_gtid_slave.test > new file mode 100644 > index 0000000..8ef8745 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave.test > @@ -0,0 +1,78 @@ > +# > +# Test Galera as a slave to a MariaDB master using GTIDs > +# > +# suite/galera/galera_2nodes_as_slave.cnf describes the setup of the > nodes > +# suite/galera/t/galera_as_slave_gtid.cnf has the GTID options > +# > +# In addition to performing DDL and DML, we check that the gtid of the > master is preserved inside the cluster > +# > + > +--source include/have_innodb.inc > + > +# As node #1 is not a Galera node, we connect to node #2 in order to run > include/galera_cluster.inc > +--connect node_2a, 127.0.0.1, root, , test, $NODE_MYPORT_2 > +--source include/galera_cluster.inc > + > +--connection node_2 > +--disable_query_log > +--eval CHANGE MASTER TO MASTER_HOST='127.0.0.1', MASTER_USER='root', > MASTER_PORT=$NODE_MYPORT_1; > +--enable_query_log > +START SLAVE; > + > +--connection node_1 > +CREATE TABLE t1 (f1 INTEGER PRIMARY KEY) ENGINE=InnoDB; > +CREATE TABLE t2 (f1 INTEGER PRIMARY KEY) ENGINE=InnoDB; > +INSERT INTO t1 VALUES(1); > + > +#multi stmt trans > +begin; > +insert into t2 values(21); > +insert into t2 values(22); > +commit; > + > +SELECT @@global.gtid_binlog_state; > + > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 1 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't1'; > +--source include/wait_condition.inc > + > +--let $wait_condition = SELECT COUNT(*) = 1 FROM t1; > +--source include/wait_condition.inc > +--sleep 1 > +INSERT INTO t1 VALUES(2); > +INSERT INTO t1 VALUES(3); > +SELECT @@global.gtid_binlog_state; > + > +--connect node_3, 127.0.0.1, root, , test, $NODE_MYPORT_3 > +--let $wait_condition = SELECT COUNT(*) = 3 FROM t1; > +--source include/wait_condition.inc > + > +INSERT INTO t1 VALUES(4); > +SELECT @@global.gtid_binlog_state; > + > +--connection node_1 > +DROP TABLE t1,t2; > +reset master; > +# > +# Unfortunately without the sleep below the following statement fails > with "query returned no rows", which > +# is difficult to understand given that it is an aggregate query. A > "query execution was interrupted" > +# warning is also reported by MTR, which is also weird. > +# > + > +--sleep 1 > + > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't1'; > +--source include/wait_condition.inc > + > +--connection node_3 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't1'; > +--source include/wait_condition.inc > + > +--connection node_2 > +STOP SLAVE; > +RESET SLAVE ALL; > +reset master; > + > +--connection node_3 > +reset master; > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.cnf > b/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.cnf > new file mode 100644 > index 0000000..0b4a776 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.cnf > @@ -0,0 +1,19 @@ > +!include ../galera_2nodes_as_slave.cnf > + > +[mysqld] > +log-bin=mysqld-bin > +log-slave-updates > +binlog-format=ROW > +wsrep_sst_method=xtrabackup > + > +[mysqld.1] > +gtid-domain-id=1 > +[mysqld.2] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > +[mysqld.3] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > + > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.test > b/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.test > new file mode 100644 > index 0000000..b507698 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.test > @@ -0,0 +1,208 @@ > +# > +# Test Galera as a slave to a MariaDB master using GTIDs > +# > +# suite/galera/galera_2nodes_as_slave.cnf describes the setup of the > nodes > +# > +# In addition to performing DDL and DML, we check that the gtid of the > master is preserved inside the cluster > +# > + > +--source include/big_test.inc > +--source include/have_innodb.inc > +# As node #1 is not a Galera node, we connect to node #2 in order to run > include/galera_cluster.inc > +--connect node_2a, 127.0.0.1, root, , test, $NODE_MYPORT_2 > +--source include/galera_cluster.inc > + > +--echo #Connection 2 > +--connection node_2 > +--disable_query_log > +--eval CHANGE MASTER TO MASTER_HOST='127.0.0.1', MASTER_USER='root', > MASTER_PORT=$NODE_MYPORT_1,master_use_gtid=slave_pos; > +--enable_query_log > +START SLAVE; > +--source include/wait_for_slave_to_start.inc > + > +--echo #Connection 1 > +--connection node_1 > +CREATE TABLE t2 (f1 INTEGER PRIMARY KEY, f2 int unique) ENGINE=InnoDB; > +INSERT INTO t2 VALUES(1,11); > +INSERT INTO t2 VALUES(2,22); > +INSERT INTO t2 VALUES(3,33); > + > +SELECT @@global.gtid_binlog_state; > +--source include/save_master_gtid.inc > + > +--echo #Connection 2 > +--connection node_2 > +--source include/sync_with_master_gtid.inc > +SELECT @@global.gtid_binlog_state; > +INSERT INTO t2 VALUES(4,44); > +INSERT INTO t2 VALUES(5,55); > +INSERT INTO t2 VALUES(6,66); > +SELECT @@global.gtid_binlog_state; > + > +--echo #Connection 3 > +--connect node_3, 127.0.0.1, root, , test, $NODE_MYPORT_3 > +--let $wait_condition = SELECT COUNT(*) = 1 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME= 't2'; > +--source include/wait_condition.inc > +--let $wait_condition = SELECT COUNT(*) = 6 FROM t2; > +--source include/wait_condition.inc > + > +INSERT INTO t2 VALUES(7,77); > +INSERT INTO t2 VALUES(8,88); > +SELECT @@global.gtid_binlog_state; > + > +#Perform SST > +--echo #Connection 1 > +--connection node_1 > +CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_before'); > +INSERT INTO t1 VALUES ('node1_committed_before'); > +COMMIT; > +--source include/save_master_gtid.inc > + > +--echo #Connection 2 > +--connection node_2 > +--source include/sync_with_master_gtid.inc > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node2_committed_before'); > +INSERT INTO t1 VALUES ('node2_committed_before'); > +COMMIT; > + > +--echo #Connection 3 > +--connection node_3 > +--let $wait_condition = SELECT COUNT(*) = 1 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME= 't1'; > +--source include/wait_condition.inc > +--let $wait_condition = SELECT COUNT(*) = 4 FROM t1; > +--source include/wait_condition.inc > +--let $node_1= node_2 > +--let $node_2= node_3 > +--source include/auto_increment_offset_save.inc > +--echo Shutting down server ... > +--source include/shutdown_mysqld.inc > + > + > +--echo #Connection 2 > +--connection node_2 > +--let $wait_condition = SELECT VARIABLE_VALUE = 1 FROM > INFORMATION_SCHEMA.GLOBAL_STATUS WHERE VARIABLE_NAME = > 'wsrep_cluster_size' > +--source include/wait_condition.inc > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_during'); > +INSERT INTO t1 VALUES ('node1_committed_during'); > +COMMIT; > + > +--echo #Connection 3 > +--connection node_3 > +--echo Starting server ... > +--source include/start_mysqld.inc > +--source include/wait_until_ready.inc > +--source include/auto_increment_offset_restore.inc > + > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node3_committed_after'); > +INSERT INTO t1 VALUES ('node3_committed_after'); > +COMMIT; > + > +--echo #Connection 2 > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 8 FROM t1; > +--source include/wait_condition.inc > +Select * from t1 order by f1; > + > +--echo #Connection 3 > +--connection node_3 > +Select * from t1 order by f1; > + > +#SST Done > +--sleep 1 > +--echo #Connection 2 > +--connection node_2 > +SELECT @@global.gtid_binlog_state; > + > +--echo #Connection 3 > +--connection node_3 > +SELECT @@global.gtid_binlog_state; > + > +--echo #Connection 1 > +--connection node_1 > +SET AUTOCOMMIT=ON; > +#drop table t1; > +#CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > + > +--echo #Connection 2 > +--connection node_2 > +SET AUTOCOMMIT=ON; > +--echo #Connection 3 > +--connection node_3 > +SET AUTOCOMMIT=ON; > + > +# > +#stop slave on node 2 > +--echo #Connection 2 > +--connection node_2 > +STOP slave; > +--source include/wait_for_slave_to_stop.inc > +--sleep 1 > +INSERT INTO t1 VALUES ('node2_slave_stoped'); > + > +--echo #Connection 1 > +--connection node_1 > +INSERT INTO t1 VALUES ('node1_normal_entry'); > +--source include/save_master_gtid.inc > + > +#start slave > +--echo #Connection 2 > +--connection node_2 > +INSERT INTO t1 VALUES ('node2_slave_stoped_inserted'); > +start slave; > +--source include/wait_for_slave_to_start.inc > +--source include/sync_with_master_gtid.inc > +INSERT INTO t1 VALUES ('node2_slave_started'); > +SELECT count(*) from t1; > +SELECT @@global.gtid_binlog_state; > + > +--echo #Connection 3 > +--connection node_3 > +--let $wait_condition = SELECT COUNT(*) = 12 FROM t1; > +--source include/wait_condition.inc > +SELECT count(*) from t1; > +SELECT @@global.gtid_binlog_state; > + > +--echo #Connection 1 > +--connection node_1 > +DROP TABLE t2,t1; > + > +# Unfortunately without the sleep below the following statement fails > with "query returned no rows", which > +# is difficult to understand given that it is an aggregate query. A > "query execution was interrupted" > +# warning is also reported by MTR, which is also weird. > +# > + > +--sleep 3 > + > +--echo #Connection 2 > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't2'; > +--source include/wait_condition.inc > + > +--echo #Connection 3 > +--connection node_3 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't1'; > +--source include/wait_condition.inc > + > +--echo #Connection 2 > +--connection node_2 > +STOP SLAVE; > +RESET SLAVE; > +reset master; > + > +--echo #Connection 3 > +--connection node_3 > +reset master; > + > +--echo #Connection 1 > +--connection node_1 > +reset master; > + > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.cnf > b/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.cnf > new file mode 100644 > index 0000000..bb9c8e8 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.cnf > @@ -0,0 +1,18 @@ > +!include ../galera_2nodes_as_slave.cnf > + > +[mysqld] > +log-bin=mysqld-bin > +log-slave-updates > +binlog-format=ROW > +wsrep_sst_method=rsync > +[mysqld.1] > +gtid-domain-id=1 > +[mysqld.2] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > +[mysqld.3] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > + > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.test > b/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.test > new file mode 100644 > index 0000000..3fe94ad > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.test > @@ -0,0 +1,207 @@ > +# > +# Test Galera as a slave to a MariaDB master using GTIDs > +# > +# suite/galera/galera_2nodes_as_slave.cnf describes the setup of the > nodes > +# > +# In addition to performing DDL and DML, we check that the gtid of the > master is preserved inside the cluster > +# > + > +--source include/big_test.inc > +--source include/have_innodb.inc > +# As node #1 is not a Galera node, we connect to node #2 in order to run > include/galera_cluster.inc > +--connect node_2a, 127.0.0.1, root, , test, $NODE_MYPORT_2 > +--source include/galera_cluster.inc > + > +--echo #Connection 2 > +--connection node_2 > +--disable_query_log > +--eval CHANGE MASTER TO MASTER_HOST='127.0.0.1', MASTER_USER='root', > MASTER_PORT=$NODE_MYPORT_1,master_use_gtid=slave_pos; > +--enable_query_log > +START SLAVE; > +--sleep 1 > + > + > +--echo #Connection 1 > +--connection node_1 > +CREATE TABLE t2 (f1 INTEGER PRIMARY KEY, f2 int unique) ENGINE=InnoDB; > +INSERT INTO t2 VALUES(1,11); > +INSERT INTO t2 VALUES(2,22); > +INSERT INTO t2 VALUES(3,33); > + > +SELECT @@global.gtid_binlog_state; > +--source include/save_master_gtid.inc > + > +--echo #Connection 2 > +--connection node_2 > +--source include/sync_with_master_gtid.inc > +SELECT @@global.gtid_binlog_state; > + > +INSERT INTO t2 VALUES(4,44); > +INSERT INTO t2 VALUES(5,55); > +INSERT INTO t2 VALUES(6,66); > +SELECT @@global.gtid_binlog_state; > + > +--echo #Connection 3 > +--connect node_3, 127.0.0.1, root, , test, $NODE_MYPORT_3 > +--let $wait_condition = SELECT COUNT(*) = 1 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME= 't2'; > +--source include/wait_condition.inc > +--let $wait_condition = SELECT COUNT(*) = 6 FROM t2; > +--source include/wait_condition.inc > + > +INSERT INTO t2 VALUES(7,77); > +INSERT INTO t2 VALUES(8,88); > +SELECT @@global.gtid_binlog_state; > + > +#Perform SST > +--echo #Connection 1 > +--connection node_1 > +CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_before'); > +INSERT INTO t1 VALUES ('node1_committed_before'); > +COMMIT; > +--source include/save_master_gtid.inc > + > +--echo #Connection 2 > +--connection node_2 > +--source include/sync_with_master_gtid.inc > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node2_committed_before'); > +INSERT INTO t1 VALUES ('node2_committed_before'); > +COMMIT; > + > +--echo #Connection 3 > +--connection node_3 > +--let $wait_condition = SELECT COUNT(*) = 1 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME= 't1'; > +--source include/wait_condition.inc > +--let $wait_condition = SELECT COUNT(*) = 4 FROM t1; > +--source include/wait_condition.inc > +--let $node_1= node_2 > +--let $node_2= node_3 > +--source include/auto_increment_offset_save.inc > +--echo Shutting down server ... > +--source include/shutdown_mysqld.inc > + > + > +--echo #Connection 2 > +--connection node_2 > +--let $wait_condition = SELECT VARIABLE_VALUE = 1 FROM > INFORMATION_SCHEMA.GLOBAL_STATUS WHERE VARIABLE_NAME = > 'wsrep_cluster_size' > +--source include/wait_condition.inc > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_during'); > +INSERT INTO t1 VALUES ('node1_committed_during'); > +COMMIT; > + > +--echo #Connection 3 > +--connection node_3 > +--echo Starting server ... > +--source include/start_mysqld.inc > +--source include/wait_until_ready.inc > +--source include/auto_increment_offset_restore.inc > + > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node3_committed_after'); > +INSERT INTO t1 VALUES ('node3_committed_after'); > +COMMIT; > + > +--echo #Connection 2 > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 8 FROM t1; > +--source include/wait_condition.inc > +Select * from t1 order by f1; > + > +--echo #Connection 3 > +--connection node_3 > +Select * from t1 order by f1; > + > +#SST Done > +--sleep 1 > +--echo #Connection 2 > +--connection node_2 > +SELECT @@global.gtid_binlog_state; > + > +--echo #Connection 3 > +--connection node_3 > +SELECT @@global.gtid_binlog_state; > + > +--echo #Connection 1 > +--connection node_1 > +SET AUTOCOMMIT=ON; > +#drop table t1; > +#CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > + > +--echo #Connection 2 > +--connection node_2 > +SET AUTOCOMMIT=ON; > +--echo #Connection 3 > +--connection node_3 > +SET AUTOCOMMIT=ON; > + > +# > +#stop slave on node 2 > +--echo #Connection 2 > +--connection node_2 > +STOP slave; > +--sleep 1 > +INSERT INTO t1 VALUES ('node2_slave_stoped'); > + > +--echo #Connection 1 > +--connection node_1 > +INSERT INTO t1 VALUES ('node1_normal_entry'); > +--source include/save_master_gtid.inc > + > +#start slave > +--echo #Connection 2 > +--connection node_2 > +INSERT INTO t1 VALUES ('node2_slave_stoped_inserted'); > +start slave; > +--source include/sync_with_master_gtid.inc > +INSERT INTO t1 VALUES ('node2_slave_started'); > +SELECT count(*) from t1; > +SELECT @@global.gtid_binlog_state; > + > +--echo #Connection 3 > +--connection node_3 > +--let $wait_condition = SELECT COUNT(*) = 12 FROM t1; > +--source include/wait_condition.inc > +SELECT count(*) from t1; > +SELECT @@global.gtid_binlog_state; > + > +--echo #Connection 1 > +--connection node_1 > +DROP TABLE t2,t1; > + > +# Unfortunately without the sleep below the following statement fails > with "query returned no rows", which > +# is difficult to understand given that it is an aggregate query. A > "query execution was interrupted" > +# warning is also reported by MTR, which is also weird. > +# > + > +--sleep 3 > + > +--echo #Connection 2 > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't2'; > +--source include/wait_condition.inc > + > +--echo #Connection 3 > +--connection node_3 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM > INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't1'; > +--source include/wait_condition.inc > + > +--echo #Connection 2 > +--connection node_2 > +STOP SLAVE; > +RESET SLAVE ALL; > +reset master; > + > +--echo #Connection 3 > +--connection node_3 > +reset master; > + > +--echo #Connection 1 > +--connection node_1 > +reset master; > diff --git a/mysql-test/suite/galera_3nodes/galera_2x3nodes.cnf > b/mysql-test/suite/galera_3nodes/galera_2x3nodes.cnf > new file mode 100644 > index 0000000..3f39b82 > --- /dev/null > +++ b/mysql-test/suite/galera_3nodes/galera_2x3nodes.cnf > @@ -0,0 +1,122 @@ > +# Use default setting for mysqld processes > +!include include/default_mysqld.cnf > + > +[mysqld] > +log-bin > +log-slave-updates > +binlog-format=row > +innodb-autoinc-lock-mode=2 > +default-storage-engine=innodb > +wsrep_gtid_mode=1 > +gtid_ignore_duplicates > + > +wsrep-on=1 > +wsrep-provider=(a)ENV.WSREP_PROVIDER > +wsrep_node_address=127.0.0.1 > +# enforce read-committed characteristics across the cluster > +# wsrep-causal-reads=ON > +# wsrep-sync-wait=15 > + > +[mysqld.1] > +#galera_port=(a)OPT.port > +#ist_port=(a)OPT.port > +#sst_port=(a)OPT.port > +wsrep-cluster-address='gcomm://' > +wsrep_provider_options='base_port=(a)mysqld.1.#galera_port; > evs.suspect_timeout=PT300S;evs.inactive_timeout=PT1000M; > evs.install_timeout=PT155S;evs.keepalive_period = PT100S' > + > +wsrep_sst_receive_address=127.0.0.2:@mysqld.1.#sst_port > +wsrep_node_incoming_address=127.0.0.1:@mysqld.1.port > +wsrep_sst_receive_address='127.0.0.1:@mysqld.1.#sst_port' > + > +[mysqld.2] > +#galera_port=(a)OPT.port > +#ist_port=(a)OPT.port > +#sst_port=(a)OPT.port > +wsrep_cluster_address='gcomm://127.0.0.1:@mysqld.1.#galera_port' > +wsrep_provider_options='base_port=(a)mysqld.2.#galera_port; > evs.suspect_timeout=PT300S;evs.inactive_timeout=PT1000M; > evs.install_timeout=PT155S;evs.keepalive_period = PT100S' > + > +wsrep_sst_receive_address=127.0.0.2:@mysqld.2.#sst_port > +wsrep_node_incoming_address=127.0.0.1:@mysqld.2.port > +wsrep_sst_receive_address='127.0.0.1:@mysqld.2.#sst_port' > + > +[mysqld.3] > +#galera_port=(a)OPT.port > +#ist_port=(a)OPT.port > +#sst_port=(a)OPT.port > +wsrep_cluster_address='gcomm://127.0.0.1:@mysqld.1.#galera_port' > +wsrep_provider_options='base_port=(a)mysqld.3.#galera_port; > evs.suspect_timeout=PT300S;evs.inactive_timeout=PT1000M; > evs.install_timeout=PT155S;evs.keepalive_period = PT100S' > + > +wsrep_sst_receive_address=127.0.0.2:@mysqld.3.#sst_port > +wsrep_node_incoming_address=127.0.0.1:@mysqld.3.port > +wsrep_sst_receive_address='127.0.0.1:@mysqld.3.#sst_port' > + > + > +[mysqld.4] > +wsrep_cluster_name=cluster2 > +#galera_port=(a)OPT.port > +#ist_port=(a)OPT.port > +#sst_port=(a)OPT.port > + > +wsrep-cluster-address='gcomm://' > +wsrep_provider_options='base_port=(a)mysqld.4.#galera_port; > evs.suspect_timeout=PT300S;evs.inactive_timeout=PT1000M; > evs.install_timeout=PT155S;evs.keepalive_period = PT100S' > + > +wsrep_sst_receive_address=127.0.0.2:@mysqld.4.#sst_port > +wsrep_node_incoming_address=127.0.0.1:@mysqld.4.port > +wsrep_sst_receive_address='127.0.0.1:@mysqld.4.#sst_port' > + > +[mysqld.5] > +wsrep_cluster_name=cluster2 > +#galera_port=(a)OPT.port > +#ist_port=(a)OPT.port > +#sst_port=(a)OPT.port > +wsrep_cluster_address='gcomm://127.0.0.1:@mysqld.4.#galera_port' > +wsrep_provider_options='base_port=(a)mysqld.5.#galera_port; > evs.suspect_timeout=PT300S;evs.inactive_timeout=PT1000M; > evs.install_timeout=PT155S;evs.keepalive_period = PT100S' > + > +wsrep_sst_receive_address=127.0.0.2:@mysqld.5.#sst_port > +wsrep_node_incoming_address=127.0.0.1:@mysqld.5.port > +wsrep_sst_receive_address='127.0.0.1:@mysqld.5.#sst_port' > + > +[mysqld.6] > +wsrep_cluster_name=cluster2 > +#galera_port=(a)OPT.port > +#ist_port=(a)OPT.port > +#sst_port=(a)OPT.port > +wsrep_cluster_address='gcomm://127.0.0.1:@mysqld.4.#galera_port' > +wsrep_provider_options='base_port=(a)mysqld.6.#galera_port; > evs.suspect_timeout=PT300S;evs.inactive_timeout=PT1000M; > evs.install_timeout=PT155S;evs.keepalive_period = PT100S' > + > +wsrep_sst_receive_address=127.0.0.2:@mysqld.6.#sst_port > +wsrep_node_incoming_address=127.0.0.1:@mysqld.6.port > +wsrep_sst_receive_address='127.0.0.1:@mysqld.6.#sst_port' > + > +[ENV] > +NODE_MYPORT_1= @mysqld.1.port > +NODE_MYSOCK_1= @mysqld.1.socket > + > +NODE_MYPORT_2= @mysqld.2.port > +NODE_MYSOCK_2= @mysqld.2.socket > + > +NODE_MYPORT_3= @mysqld.3.port > +NODE_MYSOCK_3= @mysqld.3.socket > + > +NODE_MYPORT_4= @mysqld.4.port > +NODE_MYSOCK_4= @mysqld.4.socket > + > +NODE_MYPORT_5= @mysqld.5.port > +NODE_MYSOCK_5= @mysqld.5.socket > + > +NODE_MYPORT_6= @mysqld.6.port > +NODE_MYSOCK_6= @mysqld.6.socket > + > +NODE_GALERAPORT_1= @mysqld.1.#galera_port > +NODE_GALERAPORT_2= @mysqld.2.#galera_port > +NODE_GALERAPORT_3= @mysqld.3.#galera_port > +NODE_GALERAPORT_4= @mysqld.4.#galera_port > +NODE_GALERAPORT_5= @mysqld.5.#galera_port > +NODE_GALERAPORT_6= @mysqld.6.#galera_port > + > +NODE_SSTPORT_1= @mysqld.1.#sst_port > +NODE_SSTPORT_2= @mysqld.2.#sst_port > +NODE_SSTPORT_3= @mysqld.3.#sst_port > +NODE_SSTPORT_4= @mysqld.4.#sst_port > +NODE_SSTPORT_5= @mysqld.5.#sst_port > +NODE_SSTPORT_6= @mysqld.6.#sst_port > diff --git a/mysql-test/suite/galera_3nodes/r/galera_gtid_2_cluster.result > b/mysql-test/suite/galera_3nodes/r/galera_gtid_2_cluster.result > new file mode 100644 > index 0000000..52c82ab > --- /dev/null > +++ b/mysql-test/suite/galera_3nodes/r/galera_gtid_2_cluster.result > @@ -0,0 +1,117 @@ > +cluster 1 node 1 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > +Variable_name Value > +wsrep_cluster_size 3 > +cluster 1 node 2 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > +Variable_name Value > +wsrep_cluster_size 3 > +cluster 1 node 3 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > +Variable_name Value > +wsrep_cluster_size 3 > +cluster 2 node 1 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > +Variable_name Value > +wsrep_cluster_size 3 > +cluster 2 node 2 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > +Variable_name Value > +wsrep_cluster_size 3 > +cluster 2 node 3 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > +Variable_name Value > +wsrep_cluster_size 3 > +change master to master_host='127.0.0.1', master_user='root', > master_port=16003, master_use_gtid=current_pos, ignore_server_ids=(12,13);; > +start slave; > +include/wait_for_slave_to_start.inc > +select @@gtid_binlog_state; > +@@gtid_binlog_state > + > +select @@gtid_slave_pos; > +@@gtid_slave_pos > + > +change master to master_host='127.0.0.1', master_user='root', > master_port=16000, master_use_gtid=current_pos, ignore_server_ids=(22,23);; > +start slave; > +include/wait_for_slave_to_start.inc > +select @@gtid_binlog_state; > +@@gtid_binlog_state > + > +select @@gtid_slave_pos; > +@@gtid_slave_pos > + > +cluster 1 node 1 > +create table t1 (cluster_domain_id int ,node_server_id int, seq_no int); > +insert into t1 values (1, 11, 2); > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-11-2 > +#wait for sync cluster 1 and 2 > +include/save_master_gtid.inc > +include/sync_with_master_gtid.inc > +cluster 2 node 1 > +insert into t1 values (2, 21, 1); > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-11-2,2-21-1 > +select * from t1; > +cluster_domain_id node_server_id seq_no > +1 11 2 > +2 21 1 > +#wait for sync cluster 1 and 2 > +include/save_master_gtid.inc > +include/sync_with_master_gtid.inc > +cluster 1 node 2 > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-11-2,2-21-1 > +insert into t1 values (1, 12, 3); > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-11-2,1-12-3,2-21-1 > +#wait for sync cluster 1 and 2 > +include/save_master_gtid.inc > +include/sync_with_master_gtid.inc > +cluster 1 node 3 > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-11-2,1-12-3,2-21-1 > +insert into t1 values (1, 13, 4); > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-12-3,1-11-2,1-13-4,2-21-1 > +#wait for sync cluster 1 and 2 > +include/save_master_gtid.inc > +include/sync_with_master_gtid.inc > +cluster 2 node 2 > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-12-3,1-11-2,1-13-4,2-21-1 > +insert into t1 values (2, 22, 2); > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-12-3,1-11-2,1-13-4,2-21-1,2-22-2 > +#wait for sync cluster 2 and 1 > +include/save_master_gtid.inc > +include/sync_with_master_gtid.inc > +cluster 2 node 3 > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-12-3,1-11-2,1-13-4,2-21-1,2-22-2 > +insert into t1 values (2, 23, 3); > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-12-3,1-11-2,1-13-4,2-21-1,2-22-2,2-23-3 > +#wait for sync cluster 2 and 1 > +include/save_master_gtid.inc > +include/sync_with_master_gtid.inc > +cluster 1 node 1 > +select @@gtid_binlog_state; > +@@gtid_binlog_state > +1-12-3,1-11-2,1-13-4,2-21-1,2-22-2,2-23-3 > +drop table t1; > +stop slave; > +change master to master_use_gtid=no, ignore_server_ids=(); > +cluster 2 node 1 > +stop slave; > +change master to master_use_gtid=no, ignore_server_ids=(); > diff --git a/mysql-test/suite/galera_3nodes/t/galera_gtid_2_cluster.cnf > b/mysql-test/suite/galera_3nodes/t/galera_gtid_2_cluster.cnf > new file mode 100644 > index 0000000..dc5535e > --- /dev/null > +++ b/mysql-test/suite/galera_3nodes/t/galera_gtid_2_cluster.cnf > @@ -0,0 +1,28 @@ > +# We need a dedicated .cnf file, even if empty, in order to force this > test to run > +# alone on a freshly started cluster. Otherwise there are adverse > interactions with > +# following tests such as galera_3nodes.galera_var_dirty_reads2 > + > +!include ../galera_2x3nodes.cnf > +[mysqld.1] > +wsrep_gtid_domain_id=1 > +server-id=11 > + > +[mysqld.2] > +wsrep_gtid_domain_id=1 > +server-id=12 > + > +[mysqld.3] > +wsrep_gtid_domain_id=1 > +server-id=13 > + > +[mysqld.4] > +wsrep_gtid_domain_id=2 > +server-id=21 > + > +[mysqld.5] > +wsrep_gtid_domain_id=2 > +server-id=22 > + > +[mysqld.6] > +wsrep_gtid_domain_id=2 > +server-id=23 > diff --git a/mysql-test/suite/galera_3nodes/t/galera_gtid_2_cluster.test > b/mysql-test/suite/galera_3nodes/t/galera_gtid_2_cluster.test > new file mode 100644 > index 0000000..4c9e611 > --- /dev/null > +++ b/mysql-test/suite/galera_3nodes/t/galera_gtid_2_cluster.test > @@ -0,0 +1,143 @@ > +# > +# This test creates 2x 3 nodes galera cluster. > +# The whole test case > +# A <-> B <-> C {Galera cluster 1} > +# | {Circular Async replication} > +# D <-> E <-> F {Galera cluster 2} > +# We will write on any random node to see if gtid is consitent or not > +# Then we will kill node D and set up the replication between A and E > +# To see whether fail over works or not. > + > +--source include/big_test.inc > +--source include/galera_cluster.inc > +--source include/have_innodb.inc > + > +--connection node_1 > +--echo cluster 1 node 1 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > + > +--connection node_2 > +--echo cluster 1 node 2 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > + > +--connect node_3, 127.0.0.1, root, , test, $NODE_MYPORT_3 > +--connection node_3 > +--echo cluster 1 node 3 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > + > +--connect node_4, 127.0.0.1, root, , test, $NODE_MYPORT_4 > +--connection node_4 > +--echo cluster 2 node 1 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > + > +--connect node_5, 127.0.0.1, root, , test, $NODE_MYPORT_5 > +--connection node_5 > +--echo cluster 2 node 2 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > + > +--connect node_6, 127.0.0.1, root, , test, $NODE_MYPORT_6 > +--connection node_6 > +--echo cluster 2 node 3 > +SHOW STATUS LIKE 'wsrep_cluster_size'; > + > +--connection node_1 > +--eval change master to master_host='127.0.0.1', master_user='root', > master_port=$NODE_MYPORT_4, master_use_gtid=current_pos, > ignore_server_ids=(12,13); > +start slave; > +--source include/wait_for_slave_to_start.inc > +select @@gtid_binlog_state; > +select @@gtid_slave_pos; > +#--query_vertical SHOW SLAVE STATUS; > + > +--connection node_4 > +--eval change master to master_host='127.0.0.1', master_user='root', > master_port=$NODE_MYPORT_1, master_use_gtid=current_pos, > ignore_server_ids=(22,23); > +start slave; > +--source include/wait_for_slave_to_start.inc > +select @@gtid_binlog_state; > +select @@gtid_slave_pos; > +#--query_vertical SHOW SLAVE STATUS; > + > +--echo cluster 1 node 1 > +--connection node_1 > +create table t1 (cluster_domain_id int ,node_server_id int, seq_no int); > +insert into t1 values (1, 11, 2); > +select @@gtid_binlog_state; > + > +--echo #wait for sync cluster 1 and 2 > +--connection node_1 > +--source include/save_master_gtid.inc > +--connection node_4 > +--source include/sync_with_master_gtid.inc > + > +--echo cluster 2 node 1 > +--connection node_4 > +insert into t1 values (2, 21, 1); > +select @@gtid_binlog_state; > +select * from t1; > + > +--echo #wait for sync cluster 1 and 2 > +--connection node_1 > +--source include/save_master_gtid.inc > +--connection node_4 > +--source include/sync_with_master_gtid.inc > + > + > +--echo cluster 1 node 2 > +--connection node_2 > +select @@gtid_binlog_state; > +insert into t1 values (1, 12, 3); > +select @@gtid_binlog_state; > + > +--echo #wait for sync cluster 1 and 2 > +--connection node_1 > +--source include/save_master_gtid.inc > +--connection node_4 > +--source include/sync_with_master_gtid.inc > + > +--echo cluster 1 node 3 > +--connection node_3 > +select @@gtid_binlog_state; > +insert into t1 values (1, 13, 4); > +select @@gtid_binlog_state; > + > +--echo #wait for sync cluster 1 and 2 > +--connection node_1 > +--source include/save_master_gtid.inc > +--connection node_4 > +--source include/sync_with_master_gtid.inc > + > +--echo cluster 2 node 2 > +--connection node_5 > +select @@gtid_binlog_state; > +insert into t1 values (2, 22, 2); > +select @@gtid_binlog_state; > + > +--echo #wait for sync cluster 2 and 1 > +--connection node_4 > +--source include/save_master_gtid.inc > +--connection node_1 > +--source include/sync_with_master_gtid.inc > + > +--echo cluster 2 node 3 > +--connection node_6 > +select @@gtid_binlog_state; > +insert into t1 values (2, 23, 3); > +select @@gtid_binlog_state; > + > +--echo #wait for sync cluster 2 and 1 > +--connection node_4 > +--source include/save_master_gtid.inc > +--connection node_1 > +--source include/sync_with_master_gtid.inc > + > + > +--echo cluster 1 node 1 > +--connection node_1 > +select @@gtid_binlog_state; > +drop table t1; > +stop slave; > +change master to master_use_gtid=no, ignore_server_ids=(); > + > +--echo cluster 2 node 1 > +--connection node_4 > +stop slave; > +change master to master_use_gtid=no, ignore_server_ids=(); > diff --git a/sql/log.cc b/sql/log.cc > index 07d4074..bb8b1d3 100644 > --- a/sql/log.cc > +++ b/sql/log.cc > @@ -5544,7 +5544,36 @@ THD::binlog_start_trans_and_stmt() > cache_mngr->trx_cache.get_prev_position() == MY_OFF_T_UNDEF) > { > this->binlog_set_stmt_begin(); > - if (in_multi_stmt_transaction_mode()) > + bool mstmt_mode= in_multi_stmt_transaction_mode(); > +#ifdef WITH_WSREP > + /* Write Gtid > + Get domain id only when gtid mode is set > + If this event is replicate through a master then , > + we will forward the same gtid another nodes > + We have to do this only one time in mysql transaction. > + Since this function is called multiple times , We will check for > + ha_info->is_started() > + */ > + Ha_trx_info *ha_info; > + ha_info= this->ha_data[binlog_hton->slot].ha_info + (mstmt_mode ? 1 > : 0); > + > + if (!ha_info->is_started() && wsrep_gtid_mode && > this->variables.gtid_seq_no) > + { > + binlog_cache_mngr *const cache_mngr= > + (binlog_cache_mngr*) thd_get_ha_data(this, binlog_hton); > + > + IO_CACHE *file= > + cache_mngr->get_binlog_cache_log(use_trans_cache(this, true)); > + Log_event_writer writer(file); > + Gtid_log_event gtid_event(this, this->variables.gtid_seq_no, > + this->variables.gtid_domain_id, > + true, LOG_EVENT_SUPPRESS_USE_F, > + true, 0); > + gtid_event.server_id= this->variables.server_id; > + writer.write(&gtid_event); > + } > +#endif > + if (mstmt_mode) > trans_register_ha(this, TRUE, binlog_hton); > trans_register_ha(this, FALSE, binlog_hton); > /* > @@ -5826,7 +5855,7 @@ MYSQL_BIN_LOG::write_gtid_event(THD *thd, bool > standalone, > DBUG_PRINT("enter", ("standalone: %d", standalone)); > > #ifdef WITH_WSREP > - if (WSREP(thd) && thd->wsrep_trx_meta.gtid.seqno != -1 && > wsrep_gtid_mode) > + if (WSREP(thd) && thd->wsrep_trx_meta.gtid.seqno != -1 && > wsrep_gtid_mode && !thd->variables.gtid_seq_no) > { > domain_id= wsrep_gtid_domain_id; > } else { > @@ -6840,6 +6869,24 @@ int MYSQL_BIN_LOG::write_cache(THD *thd, IO_CACHE > *cache) > mysql_mutex_assert_owner(&LOCK_log); > if (reinit_io_cache(cache, READ_CACHE, 0, 0, 0)) > DBUG_RETURN(ER_ERROR_ON_WRITE); > + uchar *read_pos= cache->read_pos; > + > +#ifdef WITH_WSREP > + /* > + In the cache we have gtid event if , below condition is true, > + then we will simply increment the read position of cache by > + the length of Gtid_log_event. > + */ > + if (wsrep_gtid_mode && thd->variables.gtid_seq_no) > + { > + uchar *head= read_pos; > + uint data_len= uint4korr(head + EVENT_LEN_OFFSET); > + uint event_type= (uchar)head[EVENT_TYPE_OFFSET]; > + if (event_type == GTID_LOG_EVENT) > + cache->read_pos+= data_len; > + } > +#endif > + > uint length= my_b_bytes_in_cache(cache), group, carry, hdr_offs; > long val; > ulong end_log_pos_inc= 0; // each event processed adds > BINLOG_CHECKSUM_LEN 2 t > diff --git a/sql/wsrep_mysqld.cc b/sql/wsrep_mysqld.cc > index bdc8491..285bb52 100644 > --- a/sql/wsrep_mysqld.cc > +++ b/sql/wsrep_mysqld.cc > @@ -1227,6 +1227,16 @@ int wsrep_to_buf_helper( > if (!ret && writer.write(&gtid_ev)) ret= 1; > } > #endif /* GTID_SUPPORT */ > + if (wsrep_gtid_mode && thd->variables.gtid_seq_no) > + { > + Gtid_log_event gtid_event(thd, thd->variables.gtid_seq_no, > + thd->variables.gtid_domain_id, > + true, LOG_EVENT_SUPPRESS_USE_F, > + true, 0); > + gtid_event.server_id= thd->variables.server_id; > + if (!gtid_event.is_valid()) ret= 0; > + ret= writer.write(&gtid_event); > + } > > /* if there is prepare query, add event for it */ > if (!ret && thd->wsrep_TOI_pre_query) > > _______________________________________________ > commits mailing list > commits(a)mariadb.org > https://lists.askmonty.org/cgi-bin/mailman/listinfo/commits

1 0

[Maria-developers] Issue with commit 0f43279c / MDEV-13936
by jerome brauge 12 Dec '17

12 Dec '17

Hi All, Since some time ago, my app crashed the server (10.2.-ext) in some case (repeatable) with the following stack trace : On Linux Centos 7: ./sql/mysqld(my_print_stacktrace+0x38)[0x7f463f49a093] ./sql/mysqld(handle_fatal_signal+0x3a3)[0x7f463ed2747a] /lib64/libpthread.so.0(+0xf370)[0x7f463dffd370] ./sql/mysqld(_ZN13Explain_query10get_selectEj+0x40)[0x7f463ebf4d06] ./sql/mysqld(_ZN13Explain_query8add_nodeEP12Explain_node+0x203)[0x7f463ebf4f15] ./sql/mysqld(_ZN4JOIN24save_explain_data_internEP13Explain_querybbbPKc+0x959)[0x7f463eaf3db9] ./sql/mysqld(_ZN4JOIN17save_explain_dataEP13Explain_querybbbb+0x14b)[0x7f463eabb963] ./sql/mysqld(_ZN4JOIN13build_explainEv+0xe2)[0x7f463eab3b70] ./sql/mysqld(_ZN4JOIN8optimizeEv+0xb6)[0x7f463eab3d2e] ./sql/mysqld(_ZN13st_select_lex31optimize_unflattened_subqueriesEb+0x263)[0x7f463ea57b5d] ./sql/mysqld(_ZN4JOIN31optimize_unflattened_subqueriesEv+0x24)[0x7f463ec1b746] ./sql/mysqld(_ZN4JOIN15optimize_stage2Ev+0x1cdb)[0x7f463eab7171] ./sql/mysqld(_ZN4JOIN14optimize_innerEv+0x14dd)[0x7f463eab542d] ./sql/mysqld(_ZN4JOIN8optimizeEv+0x7a)[0x7f463eab3cf2] ./sql/mysqld(_Z12mysql_selectP3THDP10TABLE_LISTjR4ListI4ItemEPS4_jP8st_orderS9_S7_S9_yP13select_resultP18st_select_lex_unitP13st_select_lex+0x263)[0x7f463eabccd2] ./sql/mysqld(_Z13handle_selectP3THDP3LEXP13select_resultm+0x16c)[0x7f463eab13e6] ./sql/mysqld(_Z21mysql_execute_commandP3THD+0x4b5b)[0x7f463ea76fb4] ./sql/mysqld(_ZN13sp_instr_stmt9exec_coreEP3THDPj+0x80)[0x7f463eeaafb0] ./sql/mysqld(_ZN13sp_lex_keeper23reset_lex_and_exec_coreEP3THDPjbP8sp_instr+0x1d1)[0x7f463eeaa549] ./sql/mysqld(_ZN13sp_instr_stmt7executeEP3THDPj+0x1b5)[0x7f463eeaabf9] ./sql/mysqld(_ZN7sp_head7executeEP3THDb+0x754)[0x7f463eea4ff6] ./sql/mysqld(_ZN7sp_head17execute_procedureEP3THDP4ListI4ItemE+0x971)[0x7f463eea74df] ./sql/mysqld(+0x645959)[0x7f463ea71959] ./sql/mysqld(_ZN12Sql_cmd_call7executeEP3THD+0x270)[0x7f463ea7240c] ./sql/mysqld(_Z21mysql_execute_commandP3THD+0x9f95)[0x7f463ea7c3ee] ./sql/mysqld(_ZN18Prepared_statement7executeEP6Stringb+0x508)[0x7f463ea9e468] ./sql/mysqld(_ZN18Prepared_statement12execute_loopEP6StringbPhS2_+0x200)[0x7f463ea9c6fa] ./sql/mysqld(+0x66e113)[0x7f463ea9a113] ./sql/mysqld(_Z19mysqld_stmt_executeP3THDPcj+0x97)[0x7f463ea99cd7] ./sql/mysqld(_Z16dispatch_command19enum_server_commandP3THDPcjbb+0xc9e)[0x7f463ea6e68d] ./sql/mysqld(_Z10do_commandP3THD+0x700)[0x7f463ea6d315] ./sql/mysqld(_Z24do_handle_one_connectionP7CONNECT+0x1ca)[0x7f463ebbe3d2] ./sql/mysqld(handle_one_connection+0x30)[0x7f463ebbe125] ./sql/mysqld(+0x1012852)[0x7f463f43e852] /lib64/libpthread.so.0(+0x7dc5)[0x7f463dff5dc5] /lib64/libc.so.6(clone+0x6d)[0x7f463c01673d] Or on Windows : Unhandled exception thrown: read access violation. Dynamic_array<Explain_select *>::at(...) returned 0xDB925000. > mysqld.exe!Explain_query::~Explain_query() Line 71 C++ mysqld.exe!Explain_query::`scalar deleting destructor'(unsigned int) C++ mysqld.exe!delete_explain_query(LEX * lex) Line 2397 C++ mysqld.exe!sp_lex_keeper::reset_lex_and_exec_core(THD * thd, unsigned int * nextp, bool open_tables, sp_instr * instr) Line 3130 C++ mysqld.exe!sp_instr_stmt::execute(THD * thd, unsigned int * nextp) Line 3266 C++ mysqld.exe!sp_head::execute(THD * thd, bool merge_da_on_success) Line 1213 C++ mysqld.exe!sp_head::execute_procedure(THD * thd, List<Item> * args) Line 2072 C++ mysqld.exe!do_execute_sp(THD * thd, sp_head * sp) Line 2907 C++ mysqld.exe!Sql_cmd_call::execute(THD * thd) Line 3147 C++ mysqld.exe!mysql_execute_command(THD * thd) Line 6223 C++ mysqld.exe!Prepared_statement::execute(String * expanded_query, bool open_cursor) Line 4682 C++ mysqld.exe!Prepared_statement::execute_loop(String * expanded_query, bool open_cursor, unsigned char * packet, unsigned char * packet_end) Line 4111 C++ mysqld.exe!mysql_stmt_execute_common(THD * thd, unsigned long stmt_id, unsigned char * packet, unsigned char * packet_end, unsigned long cursor_flags, bool bulk_op, bool read_types) Line 3132 C++ mysqld.exe!mysqld_stmt_execute(THD * thd, char * packet_arg, unsigned int packet_length) Line 3028 C++ mysqld.exe!dispatch_command(enum_server_command command, THD * thd, char * packet, unsigned int packet_length, bool is_com_multi, bool is_next_command) Line 1763 C++ mysqld.exe!do_command(THD * thd) Line 1369 C++ mysqld.exe!threadpool_process_request(THD * thd) Line 366 C++ mysqld.exe!tp_callback(TP_connection * c) Line 192 C++ mysqld.exe!tp_callback(_TP_CALLBACK_INSTANCE * instance, void * context) Line 377 C++ mysqld.exe!work_callback(_TP_CALLBACK_INSTANCE * instance, void * context, _TP_WORK * work) Line 451 C++ The sql statement involved (in a stored procedure): Query (0x7f457df65bb8): INSERT INTO qnwtc( enrqnwtc, priqnwtc, lprqnwtc, cprqnwtc, nuiqnwtc) SELECT 'PRI' ,occgtpar ,intgtpar ,ch1gtpar , NAME_CONST('wpio_ident',110325)FROM gtpara WHERE padgtpar = 'PRIQNTAC' AND numgtets = NAME_CONST('wsysets',_utf8'IFR' COLLATE 'utf8_bin')AND etagtpar = 'A' Where gtpara is a view. If the stored procedure is called alone, it works fine. I've revert commit 0f43279c (MDEV-13936) and now my app works fine again. Regards.

1 0

Re: [Maria-developers] Mdev-10715 Initial Patch
by Sachin Setiya 11 Dec '17

11 Dec '17

Hi Andrei! Thanks for the review. On Thu, Oct 12, 2017 at 11:48 PM, <andrei.elkin(a)pp.inet.fi> wrote: > Howdy, Sachin! > > I've checked the patch mostly from the high-level design pov to gather > few suggestions. Find them on the top of diff hunks (hopefully the diff > remains applicable) below indented with '*review*>'. > > > > revision-id: daf9a67ea13c5c7309f85a104e7c07ec68f65e88 (mariadb-10.1.28-10-gdaf9a67) > parent(s): 5eb666ad37ff80e5f8dc714bb68fcb1e269b797a > author: Sachin Setiya > committer: Sachin Setiya > timestamp: 2017-10-10 17:45:21 +0530 > message: > > Mdev-10715 Initial Patch > *review*> The initial patch is a very good time to write down the work > intent and its key features. Done > Could you please do that and highlight how WSREP has really > become multi-source compatible. > > > --- > mysql-test/suite/galera/r/galera_gtid_slave.result | 18 +++ > .../r/galera_gtid_slave_sst_mysqldump.result | 92 ++++++++++++ > .../galera/r/galera_gtid_slave_sst_rsync.result | 92 ++++++++++++ > mysql-test/suite/galera/t/galera_gtid_slave.cnf | 18 +++ > mysql-test/suite/galera/t/galera_gtid_slave.test | 67 +++++++++ > .../galera/t/galera_gtid_slave_sst_mysqldump.cnf | 18 +++ > .../galera/t/galera_gtid_slave_sst_mysqldump.test | 156 +++++++++++++++++++++ > .../suite/galera/t/galera_gtid_slave_sst_rsync.cnf | 19 +++ > .../galera/t/galera_gtid_slave_sst_rsync.test | 156 +++++++++++++++++++++ > sql/log.cc | 37 ++++- > sql/wsrep_mysqld.cc | 9 ++ > 11 files changed, 681 insertions(+), 1 deletion(-) > > diff --git a/mysql-test/suite/galera/r/galera_gtid_slave.result b/mysql-test/suite/galera/r/galera_gtid_slave.result > new file mode 100644 > index 0000000..7ab10c7 > --- /dev/null > +++ b/mysql-test/suite/galera/r/galera_gtid_slave.result > @@ -0,0 +1,18 @@ > +START SLAVE; > +CREATE TABLE t1 (f1 INTEGER PRIMARY KEY) ENGINE=InnoDB; > +INSERT INTO t1 VALUES(1); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-2 > +INSERT INTO t1 VALUES(2); > +INSERT INTO t1 VALUES(3); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-2,2-2-2 > +INSERT INTO t1 VALUES(4); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-2,2-2-2,2-3-3 > +DROP TABLE t1; > +STOP SLAVE; > +RESET SLAVE ALL; > diff --git a/mysql-test/suite/galera/r/galera_gtid_slave_sst_mysqldump.result b/mysql-test/suite/galera/r/galera_gtid_slave_sst_mysqldump.result > new file mode 100644 > index 0000000..8e8e2e0 > --- /dev/null > +++ b/mysql-test/suite/galera/r/galera_gtid_slave_sst_mysqldump.result > @@ -0,0 +1,92 @@ > +START SLAVE; > +CREATE TABLE t2 (f1 INTEGER PRIMARY KEY, f2 int unique) ENGINE=InnoDB; > +INSERT INTO t2 VALUES(1,11); > +INSERT INTO t2 VALUES(2,22); > +INSERT INTO t2 VALUES(3,33); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4 > +INSERT INTO t2 VALUES(4,44); > +INSERT INTO t2 VALUES(5,55); > +INSERT INTO t2 VALUES(6,66); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4,2-2-3 > +INSERT INTO t2 VALUES(7,77); > +INSERT INTO t2 VALUES(8,88); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4,2-2-3,2-3-5 > +CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_before'); > +INSERT INTO t1 VALUES ('node1_committed_before'); > +COMMIT; > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node2_committed_before'); > +INSERT INTO t1 VALUES ('node2_committed_before'); > +COMMIT; > +Shutting down server ... > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_during'); > +INSERT INTO t1 VALUES ('node1_committed_during'); > +COMMIT; > +Starting server ... > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node3_committed_after'); > +INSERT INTO t1 VALUES ('node3_committed_after'); > +COMMIT; > +Select * from t1 order by f1; > +f1 > +node1_committed_before > +node1_committed_before > +node1_committed_during > +node1_committed_during > +node2_committed_before > +node2_committed_before > +node3_committed_after > +node3_committed_after > +Select * from t1 order by f1; > +f1 > +node1_committed_before > +node1_committed_before > +node1_committed_during > +node1_committed_during > +node2_committed_before > +node2_committed_before > +node3_committed_after > +node3_committed_after > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-6,2-2-7,2-3-8 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-6,2-2-7,2-3-8 > +SET AUTOCOMMIT=ON; > +SET AUTOCOMMIT=ON; > +SET AUTOCOMMIT=ON; > +STOP slave; > +INSERT INTO t1 VALUES ('node2_slave_stoped'); > +INSERT INTO t1 VALUES ('node1_slave_stoped'); > +INSERT INTO t1 VALUES ('node2_slave_stoped_inserted'); > +start slave; > +INSERT INTO t1 VALUES ('node2_slave_started'); > +SELECT COUNT(*) from t1; > +COUNT(*) > +11 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-6,2-3-8,2-2-11 > +SELECT COUNT(*) from t1; > +COUNT(*) > +11 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-6,2-3-8,2-2-11 > +DROP TABLE t2,t1; > +STOP SLAVE; > +RESET SLAVE ALL; > diff --git a/mysql-test/suite/galera/r/galera_gtid_slave_sst_rsync.result b/mysql-test/suite/galera/r/galera_gtid_slave_sst_rsync.result > new file mode 100644 > index 0000000..9ab4601 > --- /dev/null > +++ b/mysql-test/suite/galera/r/galera_gtid_slave_sst_rsync.result > @@ -0,0 +1,92 @@ > +START SLAVE; > +CREATE TABLE t2 (f1 INTEGER PRIMARY KEY, f2 int unique) ENGINE=InnoDB; > +INSERT INTO t2 VALUES(1,11); > +INSERT INTO t2 VALUES(2,22); > +INSERT INTO t2 VALUES(3,33); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4 > +INSERT INTO t2 VALUES(4,44); > +INSERT INTO t2 VALUES(5,55); > +INSERT INTO t2 VALUES(6,66); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4,2-2-3 > +INSERT INTO t2 VALUES(7,77); > +INSERT INTO t2 VALUES(8,88); > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-4,2-2-3,2-3-5 > +CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_before'); > +INSERT INTO t1 VALUES ('node1_committed_before'); > +COMMIT; > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node2_committed_before'); > +INSERT INTO t1 VALUES ('node2_committed_before'); > +COMMIT; > +Shutting down server ... > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_during'); > +INSERT INTO t1 VALUES ('node1_committed_during'); > +COMMIT; > +Starting server ... > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node3_committed_after'); > +INSERT INTO t1 VALUES ('node3_committed_after'); > +COMMIT; > +Select * from t1 order by f1; > +f1 > +node1_committed_before > +node1_committed_before > +node1_committed_during > +node1_committed_during > +node2_committed_before > +node2_committed_before > +node3_committed_after > +node3_committed_after > +Select * from t1 order by f1; > +f1 > +node1_committed_before > +node1_committed_before > +node1_committed_during > +node1_committed_during > +node2_committed_before > +node2_committed_before > +node3_committed_after > +node3_committed_after > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-6,2-2-7,2-3-8 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-6,2-2-7,2-3-8 > +SET AUTOCOMMIT=ON; > +SET AUTOCOMMIT=ON; > +SET AUTOCOMMIT=ON; > +STOP slave; > +INSERT INTO t1 VALUES ('node2_slave_stoped'); > +INSERT INTO t1 VALUES ('node1_slave_stoped'); > +INSERT INTO t1 VALUES ('node2_slave_stoped_inserted'); > +start slave; > +INSERT INTO t1 VALUES ('node2_slave_started'); > +SELECT COUNT(*) from t1; > +COUNT(*) > +11 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-7,2-3-8,2-2-11 > +SELECT COUNT(*) from t1; > +COUNT(*) > +12 > +SELECT @@global.gtid_binlog_state; > +@@global.gtid_binlog_state > +1-1-7,2-3-8,2-2-11 > +DROP TABLE t2,t1; > +STOP SLAVE; > +RESET SLAVE ALL; > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave.cnf b/mysql-test/suite/galera/t/galera_gtid_slave.cnf > new file mode 100644 > index 0000000..409d0d1 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave.cnf > @@ -0,0 +1,18 @@ > +!include ../galera_2nodes_as_slave.cnf > + > +[mysqld] > +log-bin=mysqld-bin > +log-slave-updates > +binlog-format=ROW > + > +[mysqld.1] > +gtid-domain-id=1 > +[mysqld.2] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > +[mysqld.3] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > + > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave.test b/mysql-test/suite/galera/t/galera_gtid_slave.test > new file mode 100644 > index 0000000..3b3f6f3 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave.test > @@ -0,0 +1,67 @@ > +# > +# Test Galera as a slave to a MariaDB master using GTIDs > +# > +# suite/galera/galera_2nodes_as_slave.cnf describes the setup of the nodes > +# suite/galera/t/galera_as_slave_gtid.cnf has the GTID options > +# > +# In addition to performing DDL and DML, we check that the gtid of the master is preserved inside the cluster > +# > + > +--source include/have_innodb.inc > + > +# As node #1 is not a Galera node, we connect to node #2 in order to run include/galera_cluster.inc > +--connect node_2a, 127.0.0.1, root, , test, $NODE_MYPORT_2 > +--source include/galera_cluster.inc > + > +--connection node_2 > +--disable_query_log > +--eval CHANGE MASTER TO MASTER_HOST='127.0.0.1', MASTER_USER='root', MASTER_PORT=$NODE_MYPORT_1; > +--enable_query_log > +START SLAVE; > + > +--connection node_1 > +CREATE TABLE t1 (f1 INTEGER PRIMARY KEY) ENGINE=InnoDB; > +INSERT INTO t1 VALUES(1); > + > +SELECT @@global.gtid_binlog_state; > + > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 1 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't1'; > +--source include/wait_condition.inc > + > +--let $wait_condition = SELECT COUNT(*) = 1 FROM t1; > +--source include/wait_condition.inc > +INSERT INTO t1 VALUES(2); > +INSERT INTO t1 VALUES(3); > +SELECT @@global.gtid_binlog_state; > + > +--connect node_3, 127.0.0.1, root, , test, $NODE_MYPORT_3 > +--sleep 1 > +--let $wait_condition = SELECT COUNT(*) = 3 FROM t1; > +--source include/wait_condition.inc > + > +INSERT INTO t1 VALUES(4); > +SELECT @@global.gtid_binlog_state; > + > +--connection node_1 > +DROP TABLE t1; > + > +# > +# Unfortunately without the sleep below the following statement fails with "query returned no rows", which > +# is difficult to understand given that it is an aggregate query. A "query execution was interrupted" > +# warning is also reported by MTR, which is also weird. > +# > + > +--sleep 1 > + > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't1'; > +--source include/wait_condition.inc > + > +--connection node_3 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't1'; > +--source include/wait_condition.inc > + > +--connection node_2 > +STOP SLAVE; > +RESET SLAVE ALL; > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.cnf b/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.cnf > new file mode 100644 > index 0000000..409d0d1 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.cnf > @@ -0,0 +1,18 @@ > +!include ../galera_2nodes_as_slave.cnf > + > +[mysqld] > +log-bin=mysqld-bin > +log-slave-updates > +binlog-format=ROW > + > +[mysqld.1] > +gtid-domain-id=1 > +[mysqld.2] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > +[mysqld.3] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > + > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.test b/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.test > new file mode 100644 > index 0000000..67aad21 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave_sst_mysqldump.test > @@ -0,0 +1,156 @@ > +# > +# Test Galera as a slave to a MariaDB master using GTIDs > +# > +# suite/galera/galera_2nodes_as_slave.cnf describes the setup of the nodes > +# suite/galera/t/galera_as_slave_gtid.cnf has the GTID options > +# > +# In addition to performing DDL and DML, we check that the gtid of the master is preserved inside the cluster > +# > + > +--source include/big_test.inc > +--source include/have_innodb.inc > +# As node #1 is not a Galera node, we connect to node #2 in order to run include/galera_cluster.inc > +--connect node_2a, 127.0.0.1, root, , test, $NODE_MYPORT_2 > +--source include/galera_cluster.inc > + > +--connection node_2 > +--disable_query_log > +--eval CHANGE MASTER TO MASTER_HOST='127.0.0.1', MASTER_USER='root', MASTER_PORT=$NODE_MYPORT_1; > +--enable_query_log > +START SLAVE; > + > +--connection node_1 > +CREATE TABLE t2 (f1 INTEGER PRIMARY KEY, f2 int unique) ENGINE=InnoDB; > +INSERT INTO t2 VALUES(1,11); > +INSERT INTO t2 VALUES(2,22); > +INSERT INTO t2 VALUES(3,33); > + > +SELECT @@global.gtid_binlog_state; > + > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 1 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't2'; > +--source include/wait_condition.inc > + > +--let $wait_condition = SELECT COUNT(*) = 3 FROM t2; > +--source include/wait_condition.inc > +INSERT INTO t2 VALUES(4,44); > +INSERT INTO t2 VALUES(5,55); > +INSERT INTO t2 VALUES(6,66); > +SELECT @@global.gtid_binlog_state; > + > +--connect node_3, 127.0.0.1, root, , test, $NODE_MYPORT_3 > +--sleep 1 > +--let $wait_condition = SELECT COUNT(*) = 6 FROM t2; > +--source include/wait_condition.inc > + > +INSERT INTO t2 VALUES(7,77); > +INSERT INTO t2 VALUES(8,88); > +SELECT @@global.gtid_binlog_state; > + > +#Perform SST > +--connection node_1 > +CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_before'); > +INSERT INTO t1 VALUES ('node1_committed_before'); > +COMMIT; > + > +--connection node_2 > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node2_committed_before'); > +INSERT INTO t1 VALUES ('node2_committed_before'); > +COMMIT; > + > +--connection node_3 > +--echo Shutting down server ... > +--source include/shutdown_mysqld.inc > + > + > +--connection node_2 > +--let $wait_condition = SELECT VARIABLE_VALUE = 1 FROM INFORMATION_SCHEMA.GLOBAL_STATUS WHERE VARIABLE_NAME = 'wsrep_cluster_size' > +--source include/wait_condition.inc > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_during'); > +INSERT INTO t1 VALUES ('node1_committed_during'); > +COMMIT; > + > +--connection node_3 > +--echo Starting server ... > +--source include/start_mysqld.inc > +--source include/wait_until_ready.inc > + > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node3_committed_after'); > +INSERT INTO t1 VALUES ('node3_committed_after'); > +COMMIT; > + > +--connection node_2 > +Select * from t1 order by f1; > + > +--connection node_3 > +Select * from t1 order by f1; > + > +#SST Done > +--sleep 1 > +--connection node_2 > +SELECT @@global.gtid_binlog_state; > + > +--connection node_3 > +SELECT @@global.gtid_binlog_state; > + > +# > +#stop slave on node 2 > +--connection node_1 > +SET AUTOCOMMIT=ON; > +#drop table t1; > +#CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > + > +--connection node_2 > +SET AUTOCOMMIT=ON; > +--connection node_3 > +SET AUTOCOMMIT=ON; > + > +--connection node_2 > +STOP slave; > +--sleep 1 > +INSERT INTO t1 VALUES ('node2_slave_stoped'); > + > +--connection node_1 > +INSERT INTO t1 VALUES ('node1_slave_stoped'); > +#start slave > +--connection node_2 > +INSERT INTO t1 VALUES ('node2_slave_stoped_inserted'); > +start slave; > +INSERT INTO t1 VALUES ('node2_slave_started'); > +SELECT COUNT(*) from t1; > +SELECT @@global.gtid_binlog_state; > + > +--connection node_3 > +SELECT COUNT(*) from t1; > +SELECT @@global.gtid_binlog_state; > + > +--connection node_1 > +DROP TABLE t2,t1; > + > +# Unfortunately without the sleep below the following statement fails with "query returned no rows", which > +# is difficult to understand given that it is an aggregate query. A "query execution was interrupted" > +# warning is also reported by MTR, which is also weird. > +# > + > +--sleep 3 > + > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't2'; > +--source include/wait_condition.inc > + > +--connection node_3 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't1'; > +--source include/wait_condition.inc > + > +--connection node_2 > +STOP SLAVE; > +RESET SLAVE ALL; > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.cnf b/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.cnf > new file mode 100644 > index 0000000..2c82531 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.cnf > @@ -0,0 +1,19 @@ > +!include ../galera_2nodes_as_slave.cnf > + > +[mysqld] > +log-bin=mysqld-bin > +log-slave-updates > +binlog-format=ROW > +wsrep_sst_method=rsync > + > +[mysqld.1] > +gtid-domain-id=1 > +[mysqld.2] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > +[mysqld.3] > +gtid-domain-id=2 > +wsrep_gtid_mode=1 > +wsrep_gtid_domain_id=2 > + > diff --git a/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.test b/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.test > new file mode 100644 > index 0000000..67aad21 > --- /dev/null > +++ b/mysql-test/suite/galera/t/galera_gtid_slave_sst_rsync.test > @@ -0,0 +1,156 @@ > +# > +# Test Galera as a slave to a MariaDB master using GTIDs > +# > +# suite/galera/galera_2nodes_as_slave.cnf describes the setup of the nodes > +# suite/galera/t/galera_as_slave_gtid.cnf has the GTID options > +# > +# In addition to performing DDL and DML, we check that the gtid of the master is preserved inside the cluster > +# > + > +--source include/big_test.inc > +--source include/have_innodb.inc > +# As node #1 is not a Galera node, we connect to node #2 in order to run include/galera_cluster.inc > +--connect node_2a, 127.0.0.1, root, , test, $NODE_MYPORT_2 > +--source include/galera_cluster.inc > + > +--connection node_2 > +--disable_query_log > +--eval CHANGE MASTER TO MASTER_HOST='127.0.0.1', MASTER_USER='root', MASTER_PORT=$NODE_MYPORT_1; > +--enable_query_log > +START SLAVE; > + > +--connection node_1 > +CREATE TABLE t2 (f1 INTEGER PRIMARY KEY, f2 int unique) ENGINE=InnoDB; > +INSERT INTO t2 VALUES(1,11); > +INSERT INTO t2 VALUES(2,22); > +INSERT INTO t2 VALUES(3,33); > + > +SELECT @@global.gtid_binlog_state; > + > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 1 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't2'; > +--source include/wait_condition.inc > + > +--let $wait_condition = SELECT COUNT(*) = 3 FROM t2; > +--source include/wait_condition.inc > +INSERT INTO t2 VALUES(4,44); > +INSERT INTO t2 VALUES(5,55); > +INSERT INTO t2 VALUES(6,66); > +SELECT @@global.gtid_binlog_state; > + > +--connect node_3, 127.0.0.1, root, , test, $NODE_MYPORT_3 > +--sleep 1 > +--let $wait_condition = SELECT COUNT(*) = 6 FROM t2; > +--source include/wait_condition.inc > + > +INSERT INTO t2 VALUES(7,77); > +INSERT INTO t2 VALUES(8,88); > +SELECT @@global.gtid_binlog_state; > + > +#Perform SST > +--connection node_1 > +CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_before'); > +INSERT INTO t1 VALUES ('node1_committed_before'); > +COMMIT; > + > +--connection node_2 > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node2_committed_before'); > +INSERT INTO t1 VALUES ('node2_committed_before'); > +COMMIT; > + > +--connection node_3 > +--echo Shutting down server ... > +--source include/shutdown_mysqld.inc > + > + > +--connection node_2 > +--let $wait_condition = SELECT VARIABLE_VALUE = 1 FROM INFORMATION_SCHEMA.GLOBAL_STATUS WHERE VARIABLE_NAME = 'wsrep_cluster_size' > +--source include/wait_condition.inc > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node1_committed_during'); > +INSERT INTO t1 VALUES ('node1_committed_during'); > +COMMIT; > + > +--connection node_3 > +--echo Starting server ... > +--source include/start_mysqld.inc > +--source include/wait_until_ready.inc > + > +SET AUTOCOMMIT=OFF; > +START TRANSACTION; > +INSERT INTO t1 VALUES ('node3_committed_after'); > +INSERT INTO t1 VALUES ('node3_committed_after'); > +COMMIT; > + > +--connection node_2 > +Select * from t1 order by f1; > + > +--connection node_3 > +Select * from t1 order by f1; > + > +#SST Done > +--sleep 1 > +--connection node_2 > +SELECT @@global.gtid_binlog_state; > + > +--connection node_3 > +SELECT @@global.gtid_binlog_state; > + > +# > +#stop slave on node 2 > +--connection node_1 > +SET AUTOCOMMIT=ON; > +#drop table t1; > +#CREATE TABLE t1 (f1 CHAR(255)) ENGINE=InnoDB; > + > +--connection node_2 > +SET AUTOCOMMIT=ON; > +--connection node_3 > +SET AUTOCOMMIT=ON; > + > +--connection node_2 > +STOP slave; > +--sleep 1 > +INSERT INTO t1 VALUES ('node2_slave_stoped'); > + > +--connection node_1 > +INSERT INTO t1 VALUES ('node1_slave_stoped'); > +#start slave > +--connection node_2 > +INSERT INTO t1 VALUES ('node2_slave_stoped_inserted'); > +start slave; > +INSERT INTO t1 VALUES ('node2_slave_started'); > +SELECT COUNT(*) from t1; > +SELECT @@global.gtid_binlog_state; > + > +--connection node_3 > +SELECT COUNT(*) from t1; > +SELECT @@global.gtid_binlog_state; > + > +--connection node_1 > +DROP TABLE t2,t1; > + > +# Unfortunately without the sleep below the following statement fails with "query returned no rows", which > +# is difficult to understand given that it is an aggregate query. A "query execution was interrupted" > +# warning is also reported by MTR, which is also weird. > +# > + > +--sleep 3 > + > +--connection node_2 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't2'; > +--source include/wait_condition.inc > + > +--connection node_3 > +--let $wait_condition = SELECT COUNT(*) = 0 FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_NAME = 't1'; > +--source include/wait_condition.inc > + > +--connection node_2 > +STOP SLAVE; > +RESET SLAVE ALL; > > *review*> The better timing for the non-zero seq_no Gtid write deems to > inside THD::binlog_start_trans_and_stmt() invoked few instructions > earlier. This method is invoked once per *transaction* which > guarantees no duplication which the current chosen point seems to > concede. Perhaps you change its singature to pass the writer. Done as we discussed. this is added into binlog_statrt_trans_and_stmt() And I am checking for ha_info->is_started(). > Also I think we may write the non-zero seq_no Gtid generally, regardless > of WITH_WSREP (naturally to block doing that later at binlog commit > phase, also commented below). This commit will go into 10.1 so as we discussed , we can do this later in 10.3. > > diff --git a/sql/log.cc b/sql/log.cc > index a9f486d..53eeadf 100644 > --- a/sql/log.cc > +++ b/sql/log.cc > @@ -5655,6 +5655,23 @@ int THD::binlog_write_table_map(TABLE *table, bool is_transactional, > DBUG_RETURN(error); > } > } > +#ifdef WITH_WSREP > + /* Get domain id only when gtid mode is set > + If this event is replicate through a master then , > + we will forward the same gtid another nodes > + */ > + if (wsrep_gtid_mode && this->variables.gtid_seq_no) > + { > + Gtid_log_event gtid_event(this, this->variables.gtid_seq_no, > + this->variables.gtid_domain_id, > + true, LOG_EVENT_SUPPRESS_USE_F, > + true, 0); > + gtid_event.server_id= this->variables.server_id; > + > + if ((error= writer.write(&gtid_event))) > + DBUG_RETURN(error); > + } > +#endif > if ((error= writer.write(&the_event))) > DBUG_RETURN(error); > > @@ -5826,7 +5843,7 @@ MYSQL_BIN_LOG::write_gtid_event(THD *thd, bool standalone, > DBUG_PRINT("enter", ("standalone: %d", standalone)); > > #ifdef WITH_WSREP > - if (WSREP(thd) && thd->wsrep_trx_meta.gtid.seqno != -1 && wsrep_gtid_mode) > + if (WSREP(thd) && thd->wsrep_trx_meta.gtid.seqno != -1 && wsrep_gtid_mode && !thd->variables.gtid_seq_no) > { > domain_id= wsrep_gtid_domain_id; > } else { > > *review*> I'm suggest we check the non-zero seq_no status in > MYSQL_BIN_LOG::write_gtid_event() > > *review*>diff --git a/sql/log.cc b/sql/log.cc > *review*>index 2fcb6f6dbf6..8229426a5bd 100644 > *review*>--- a/sql/log.cc > *review*>+++ b/sql/log.cc > *review*>@@ -5867,7 +5867,6 @@ MYSQL_BIN_LOG::write_gtid_event(THD *thd, bool standalone, > *review*> err= rpl_global_gtid_binlog_state.update_with_next_gtid(domain_id, > *review*> local_server_id, &gtid); > *review*> seq_no= gtid.seq_no; > *review*>- } > *review*> if (err) > *review*> DBUG_RETURN(true); > *review*> thd->last_commit_gtid= gtid; > *review*>@@ -5880,6 +5879,7 @@ MYSQL_BIN_LOG::write_gtid_event(THD *thd, bool standalone, > *review*> DBUG_ASSERT(this == &mysql_bin_log); > *review*> if (write_event(&gtid_event)) > *review*> DBUG_RETURN(true); > *review*>+ } > *review*> status_var_add(thd->status_var.binlog_bytes_written, gtid_event.data_written); > *review*> > *review*> DBUG_RETURN(false); > > *review*> so only the zero auto-evaluated seq_no Gtid would make to the > method you're trying to patch. And the hunk is unneeded in this > case. Like above we generalize WITH_WSREP away. > ^^^^^^^ I am not sure about this hunk , I think this was mistake. I have reverted this change. > @@ -6718,6 +6735,24 @@ int MYSQL_BIN_LOG::write_cache(THD *thd, IO_CACHE *cache) > mysql_mutex_assert_owner(&LOCK_log); > if (reinit_io_cache(cache, READ_CACHE, 0, 0, 0)) > DBUG_RETURN(ER_ERROR_ON_WRITE); > + uchar *read_pos= cache->read_pos; > + > +#ifdef WITH_WSREP > + /* > + In the cache we have gtid event if , below condition is true, > + then we will simply increment the read position of cache by > + the length of Gtid_log_event. > + */ > + if (wsrep_gtid_mode && thd->variables.gtid_seq_no) > + { > + uchar *head= read_pos; > + uint data_len= uint4korr(head + EVENT_LEN_OFFSET); > + uint event_type= (uchar)head[EVENT_TYPE_OFFSET]; > + if (event_type == GTID_LOG_EVENT) > + cache->read_pos+= data_len; > + } > +#endif > + > uint length= my_b_bytes_in_cache(cache), group, carry, hdr_offs; > long val; > ulong end_log_pos_inc= 0; // each event processed adds BINLOG_CHECKSUM_LEN 2 t > > *review*> The following block is fine *but* we shall explain > what is GTID_SUPPORT, whether it actually for us. > Well , there is no such thing as GTID_SUPPORT :( , we are using ifdef without even defining gtid_support ╭─sachin@sachin-desk ~/10715/server ‹bb-10.1-10715*› ╰─$ grep -Rnie 'GTID_SUPPORT' sql/wsrep_sst.cc:290:#ifdef GTID_SUPPORT sql/wsrep_sst.cc:292:#endif /* GTID_SUPPORT */ sql/wsrep_applier.cc:130:#ifdef GTID_SUPPORT sql/wsrep_applier.cc:141:#endif /* GTID_SUPPORT */ sql/wsrep_applier.cc:309:#ifdef GTID_SUPPORT sql/wsrep_applier.cc:311:#endif /* GTID_SUPPORT */ sql/wsrep_mysqld.cc:50:#ifdef GTID_SUPPORT sql/wsrep_mysqld.cc:53:#endif /* GTID_SUPPORT */ sql/wsrep_mysqld.cc:256:#ifdef GTID_SUPPORT sql/wsrep_mysqld.cc:274:#endif /* GTID_SUPPORT */ sql/wsrep_mysqld.cc:413:#ifdef GTID_SUPPORT sql/wsrep_mysqld.cc:415:#endif /* GTID_SUPPORT */ sql/wsrep_mysqld.cc:1222:#ifdef GTID_SUPPORT sql/wsrep_mysqld.cc:1229:#endif /* GTID_SUPPORT */ sql/sql_class.h:4085:#ifdef GTID_SUPPORT sql/sql_class.h:4087:#endif /* GTID_SUPPORT */ sql/wsrep_mysqld.h:261:#ifdef GTID_SUPPORT sql/wsrep_mysqld.h:263:#endif /* GTID_SUPPORT */ > diff --git a/sql/wsrep_mysqld.cc b/sql/wsrep_mysqld.cc > index 33689a9..65751f1 100644 > --- a/sql/wsrep_mysqld.cc > +++ b/sql/wsrep_mysqld.cc > @@ -1227,6 +1227,15 @@ int wsrep_to_buf_helper( > if (!ret && writer.write(&gtid_ev)) ret= 1; > } > #endif /* GTID_SUPPORT */ > + if (wsrep_gtid_mode && thd->variables.gtid_seq_no) > + { > + Gtid_log_event gtid_event(thd, thd->variables.gtid_seq_no, > + thd->variables.gtid_domain_id, > + true, LOG_EVENT_SUPPRESS_USE_F, > + true, 0); > + gtid_event.server_id= thd->variables.server_id; > + ret= writer.write(&gtid_event); > + } > > /* if there is prepare query, add event for it */ > if (!ret && thd->wsrep_TOI_pre_query) -- Regards Sachin Setiya Software Engineer at MariaDB

1 0

[Maria-developers] Mysterious files after build
by Michal Schorm 09 Dec '17

09 Dec '17

Hello, I'd like to know more about some files, that are left after MariaDB build in Fedora. I do not ship them, however I haven't found any explanation from previous maintainers. I gathered some clues and thoughts about them. Questions are: What are those files and what they does?; Is it safe to ship them?; Should I start to ship them? /usr/bin/mysql_embedded - you (MariaDB) ship this file in client package %{_datadir}/mariadb/binary-configure - this script creates the MySQL system tables and starts the server. %{_datadir}/mariadb/magic - FS files first-bytes recoginiton? %{_datadir}/mariadb/mysql.server %{_datadir}/%{pkg_name}/mysqld_multi.server - those shouldn't be needed on systemd OSs ? /usr/bin/mytop - you (mariaDB) do not ship this file -- Michal Schorm Associate Software Engineer Core Services - Databases Team Red Hat

2 1