Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-32857

Assertion `thd->lex->sql_command == SQLCOM_XA_COMMIT || thd->lex->sql_command == SQLCOM_XA_ROLLBACK' failed in TC_LOG::run_commit_ordered

Details

    Description

      Fails in bb-10.6-andrei at commit 0d728f33b6ae806a48d995738718e2c34ee94f1e (MDEV-32830 patch, latest iteration), but not at the commit prior to that (i.e. pre-MDEV-32830). It also crashes the earlier same-branch revision 88b0738b7c61a25eaffc66dda0a5a7f850c89918.

      --source include/have_innodb.inc
      		 
      --source include/have_binlog_format_mixed.inc
      		 
      CREATE TABLE t (c INT) ENGINE=InnoDB;
      		 
      XA START 'a';
      		 
      INSERT INTO t VALUES (1);
      		 
      CREATE TEMPORARY TABLE t (d INT) ENGINE=InnoDB;
      		 
      XA END 'a';
      		 
      LOAD INDEX INTO CACHE t INDEX (PRIMARY) IGNORE LEAVES;

      Leads to:

      10.6.17 0d728f33b6ae806a48d995738718e2c34ee94f1e

      		 
      mariadbd: /test/bb-10.6-andrei_dbg/sql/log.cc:9438: void TC_LOG::run_commit_ordered(THD*, bool, THD*): Assertion `thd->lex->sql_command == SQLCOM_XA_COMMIT || thd->lex->sql_command == SQLCOM_XA_ROLLBACK' failed.

      10.6.17 0d728f33b6ae806a48d995738718e2c34ee94f1e

      		 
      Core was generated by `/test/PATCH2_MD221123-mariadb-10.6.17-linux-x86_64-dbg/bin/mariadbd --defaults-'.
      		 
      Program terminated with signal SIGABRT, Aborted.
      		 
      #0  __pthread_kill_implementation (no_tid=0, signo=6, threadid=22966355924544)
      		 
          at ./nptl/pthread_kill.c:44
      		 
      [Current thread is 1 (Thread 0x14e3457c9640 (LWP 2247225))]
      		 
      (gdb) bt
      		 
      #0  __pthread_kill_implementation (no_tid=0, signo=6, threadid=22966355924544) at ./nptl/pthread_kill.c:44
      		 
      #1  __pthread_kill_internal (signo=6, threadid=22966355924544) at ./nptl/pthread_kill.c:78
      		 
      #2  __GI___pthread_kill (threadid=22966355924544, signo=signo@entry=6) at ./nptl/pthread_kill.c:89
      		 
      #3  0x000014e34be42476 in __GI_raise (sig=sig@entry=6) at ../sysdeps/posix/raise.c:26
      		 
      #4  0x000014e34be287f3 in __GI_abort () at ./stdlib/abort.c:79
      		 
      #5  0x000014e34be2871b in __assert_fail_base (fmt=0x14e34bfdd150 "%s%s%s:%u: %s%sAssertion `%s' failed.\n%n", assertion=0x558017ef7be0 "thd->lex->sql_command == SQLCOM_XA_COMMIT || thd->lex->sql_command == SQLCOM_XA_ROLLBACK", file=0x558017ef5ae8 "/test/bb-10.6-andrei_dbg/sql/log.cc", line=9438, function=<optimized out>) at ./assert/assert.c:92
      		 
      #6  0x000014e34be39e96 in __GI___assert_fail (assertion=0x558017ef7be0 "thd->lex->sql_command == SQLCOM_XA_COMMIT || thd->lex->sql_command == SQLCOM_XA_ROLLBACK", file=0x558017ef5ae8 "/test/bb-10.6-andrei_dbg/sql/log.cc", line=9438, function=0x558017ef7ba8 "void TC_LOG::run_commit_ordered(THD*, bool, THD*)") at ./assert/assert.c:101
      		 
      #7  0x00005580175734b1 in TC_LOG::run_commit_ordered (this=this@entry=0x5580186c9960 <mysql_bin_log>, thd=0x14e308000d58, all=true, leader_thd=0x14e308000d58) at /test/bb-10.6-andrei_dbg/sql/log.cc:9438
      		 
      #8  0x000055801757ca9c in MYSQL_BIN_LOG::trx_group_commit_leader (this=this@entry=0x5580186c9960 <mysql_bin_log>, leader=leader@entry=0x14e3457c6eb0) at /test/bb-10.6-andrei_dbg/sql/log.cc:8616
      		 
      #9  0x000055801757d0a8 in MYSQL_BIN_LOG::write_transaction_to_binlog_events (this=this@entry=0x5580186c9960 <mysql_bin_log>, entry=entry@entry=0x14e3457c6eb0) at /test/bb-10.6-andrei_dbg/sql/log.cc:8194
      		 
      #10 0x000055801757d6d3 in MYSQL_BIN_LOG::write_transaction_to_binlog (this=this@entry=0x5580186c9960 <mysql_bin_log>, thd=thd@entry=0x14e308000d58, cache_mngr=cache_mngr@entry=0x14e3080609c8, end_ev=end_ev@entry=0x14e3457c7190, all=all@entry=true, using_stmt_cache=using_stmt_cache@entry=false, using_trx_cache=true, is_ro_1pc=false) at /test/bb-10.6-andrei_dbg/sql/log.cc:7791
      		 
      #11 0x000055801757d8bc in binlog_flush_cache (thd=thd@entry=0x14e308000d58, cache_mngr=cache_mngr@entry=0x14e3080609c8, end_ev=end_ev@entry=0x14e3457c7190, all=all@entry=true, using_stmt=using_stmt@entry=false, using_trx=using_trx@entry=true, is_ro_1pc=false) at /test/bb-10.6-andrei_dbg/sql/log.cc:1771
      		 
      #12 0x000055801757dd27 in binlog_rollback_flush_trx_cache (thd=thd@entry=0x14e308000d58, all=all@entry=true, cache_mngr=cache_mngr@entry=0x14e3080609c8) at /test/bb-10.6-andrei_dbg/sql/log.cc:1907
      		 
      #13 0x000055801757e287 in binlog_rollback (hton=<optimized out>, thd=0x14e308000d58, all=<optimized out>) at /test/bb-10.6-andrei_dbg/sql/log.cc:2326
      		 
      #14 0x000055801742e921 in ha_rollback_trans (thd=thd@entry=0x14e308000d58, all=all@entry=true) at /test/bb-10.6-andrei_dbg/sql/handler.cc:2278
      		 
      #15 0x00005580172d571d in trans_rollback_implicit (thd=thd@entry=0x14e308000d58) at /test/bb-10.6-andrei_dbg/sql/transaction.cc:421
      		 
      #16 0x00005580172d39ae in mysql_admin_table (thd=thd@entry=0x14e308000d58, tables=tables@entry=0x14e308014830, check_opt=check_opt@entry=0x0, operator_name=operator_name@entry=0x5580185e6100 <msg_preload_keys>, lock_type=lock_type@entry=TL_READ_NO_INSERT, org_open_for_modify=org_open_for_modify@entry=false, repair_table_use_frm=<optimized out>, extra_open_options=<optimized out>, prepare_func=<optimized out>, operator_func=<optimized out>, view_operator_func=<optimized out>, is_cmd_replicated=<optimized out>) at /test/bb-10.6-andrei_dbg/sql/sql_admin.cc:1395
      		 
      #17 0x00005580172d4181 in mysql_preload_keys (thd=thd@entry=0x14e308000d58, tables=tables@entry=0x14e308014830) at /test/bb-10.6-andrei_dbg/sql/sql_admin.cc:1539
      		 
      #18 0x0000558017170b89 in mysql_execute_command (thd=thd@entry=0x14e308000d58, is_called_from_prepared_stmt=is_called_from_prepared_stmt@entry=false) at /test/bb-10.6-andrei_dbg/sql/sql_parse.cc:4103
      		 
      #19 0x0000558017177917 in mysql_parse (thd=thd@entry=0x14e308000d58, rawbuf=<optimized out>, length=<optimized out>, parser_state=parser_state@entry=0x14e3457c81f0) at /test/bb-10.6-andrei_dbg/sql/sql_parse.cc:8051
      		 
      #20 0x0000558017179c8d in dispatch_command (command=command@entry=COM_QUERY, thd=thd@entry=0x14e308000d58, packet=packet@entry=0x14e308112759 "", packet_length=packet_length@entry=53, blocking=blocking@entry=true) at /test/bb-10.6-andrei_dbg/sql/sql_class.h:241
      		 
      #21 0x000055801717bda7 in do_command (thd=0x14e308000d58, blocking=blocking@entry=true) at /test/bb-10.6-andrei_dbg/sql/sql_parse.cc:1409
      		 
      #22 0x00005580172c015d in do_handle_one_connection (connect=<optimized out>, connect@entry=0x558019999978, put_in_cache=put_in_cache@entry=true) at /test/bb-10.6-andrei_dbg/sql/sql_connect.cc:1416
      		 
      #23 0x00005580172c0452 in handle_one_connection (arg=arg@entry=0x558019999978) at /test/bb-10.6-andrei_dbg/sql/sql_connect.cc:1318
      		 
      #24 0x000055801770e656 in pfs_spawn_thread (arg=0x55801992c7e8) at /test/bb-10.6-andrei_dbg/storage/perfschema/pfs.cc:2201
      		 
      #25 0x000014e34be94ac3 in start_thread (arg=<optimized out>) at ./nptl/pthread_create.c:442
      		 
      #26 0x000014e34bf26a40 in clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:81

      Attachments

        Issue Links

          blocks

          New Feature - A new feature of the product, which has yet to be developed. MDEV-31949 slow parallel replication of user xa

          • Major - Major loss of function.
          • Stalled

          Task - A task that needs to be done. MDEV-32830 refactor XA binlogging for better integration with BGC/replication/recovery

          • Critical - Crashes, loss of data, severe memory leak.
          • In Progress
          is caused by

          Task - A task that needs to be done. MDEV-32830 refactor XA binlogging for better integration with BGC/replication/recovery

          • Critical - Crashes, loss of data, severe memory leak.
          • In Progress

          Activity

            Ascending order - Click to sort in descending order
            Elkin Andrei Elkin added a comment -

            MDEV-32455 is the reason of the assert that successfully catches an attempt to harm
            an ACTIVE xa by a (disallowed in the current case) perpetrator query type.

            Elkin Andrei Elkin added a comment - MDEV-32455 is the reason of the assert that successfully catches an attempt to harm an ACTIVE xa by a (disallowed in the current case) perpetrator query type.
            Roel Roel Van de Paar added a comment -

            It looks like this assert was newly added in the MDEV-32830 patch, and is now being triggered. As such, this is a new regression.
            A fix could be to modify the assert until MDEV-32455 is fixed. This also ensures that testing trials are not stopped short due to this assert.

            Roel Roel Van de Paar added a comment - It looks like this assert was newly added in the MDEV-32830 patch, and is now being triggered. As such, this is a new regression. A fix could be to modify the assert until MDEV-32455 is fixed. This also ensures that testing trials are not stopped short due to this assert.
            Elkin Andrei Elkin added a comment -

            Roel, I made my best to satisfy your proposal.

            Generally though it may not be easy, and I from my side may
            give up to ask you do something with your RQG grammar to exclude a particular sequence from being generated.
            Once it's tested out why should we continue be interested in seeing it over again?

            I'd love to discuss that.
            Feel to reply on slack or zulip.

            Elkin Andrei Elkin added a comment - Roel , I made my best to satisfy your proposal. Generally though it may not be easy, and I from my side may give up to ask you do something with your RQG grammar to exclude a particular sequence from being generated. Once it's tested out why should we continue be interested in seeing it over again? I'd love to discuss that. Feel to reply on slack or zulip.
            Elkin Andrei Elkin added a comment -

            The refined assert in `960e246b29d` .

            Elkin Andrei Elkin added a comment - The refined assert in `960e246b29d` .

            People

              Elkin Andrei Elkin
              Roel Roel Van de Paar
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved:

                Git Integration

                  Error rendering 'com.xiplink.jira.git.jira_git_plugin:git-issue-webpanel'. Please contact your Jira administrators.