Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-718

LP:850647 - Assertion `! is_set()' failed in Diagnostics_area::set_ok_status() with concurrent subquery workload

        Details

        • Type: Bug
        • Status: Closed
        • Priority: Major
        • Resolution: Incomplete
        • Affects Version/s: None
        • Fix Version/s: None
        • Component/s: None
        • Labels:

          Description

          mysqld: sql_class.cc:563: void Diagnostics_area::set_ok_status(THD*, ha_rows, ulonglong, const char*): Assertion `! is_set()' failed.

/lib64/libc.so.6(__assert_fail+0xf6) [0x33b58296e6]
/home/philips/bzr/maria-5.3/sql/mysqld(Diagnostics_area::set_ok_status(THD*, unsigned long long, unsigned long long, char const*)+0x68) [0x67665c]
/home/philips/bzr/maria-5.3/sql/mysqld(my_ok(THD*, unsigned long long, unsigned long long, char const*)+0x3b) [0x585bdf]
/home/philips/bzr/maria-5.3/sql/mysqld(select_insert::send_eof()+0x5d1) [0x75dfe3]
/home/philips/bzr/maria-5.3/sql/mysqld(select_create::send_eof()+0x15) [0x75e02b]
/home/philips/bzr/maria-5.3/sql/mysqld [0x755340]
/home/philips/bzr/maria-5.3/sql/mysqld(JOIN::exec()+0x788) [0x755b30]
/home/philips/bzr/maria-5.3/sql/mysqld(mysql_select(THD*, Item***, TABLE_LIST*, unsigned int, List<Item>&, Item*, unsigned int, st_order*, st_order*, Item*, st_order*, unsigned long long, select_result*, st_select_lex_unit*, st_select_lex*)+0x3da) [0x751cde]
/home/philips/bzr/maria-5.3/sql/mysqld(handle_select(THD*, st_lex*, select_result*, unsigned long)+0x1d6) [0x758108]
/home/philips/bzr/maria-5.3/sql/mysqld(mysql_execute_command(THD*)+0x1ad2) [0x6a75ce]
/home/philips/bzr/maria-5.3/sql/mysqld(mysql_parse(THD*, char*, unsigned int, char const**)+0x2c5) [0x6af02f]
/home/philips/bzr/maria-5.3/sql/mysqld(dispatch_command(enum_server_command, THD*, char*, unsigned int)+0x9cb) [0x6afbbd]
/home/philips/bzr/maria-5.3/sql/mysqld(do_command(THD*)+0x259) [0x6b11fb]
/home/philips/bzr/maria-5.3/sql/mysqld(handle_one_connection+0x18f) [0x69ba0b]

Query (0x3aa5e20): CREATE TABLE transforms.insert_select_2012   SELECT    alias2 . `pk` AS field1 , alias1 . `col_varchar_key` AS field2 , alias1 . `col_int_key` AS field3 , alias1 . `col_int_key` AS field4 , alias1 . `col_varchar_key` AS field5 FROM ( C AS alias1 INNER JOIN ( ( D AS alias2 STRAIGHT_JOIN BB AS alias3 ON (alias3 . `col_int_key` = alias2 . `pk`  ) ) ) ON (alias3 . `col_varchar_nokey` = alias2 . `col_varchar_key`  ) ) WHERE  ( 7, 1 )  IN ( SELECT   SQ1_alias2 . `col_int_nokey` AS SQ1_field1 , SQ1_alias1 . `col_int_nokey` AS SQ1_field2 FROM ( C AS SQ1_alias1 STRAIGHT_JOIN BB AS SQ1_alias2 ON (SQ1_alias2 . `pk` = SQ1_alias1 . `pk`  ) ) WHERE  SQ1_alias1 . `col_varchar_key` < ANY ( SELECT   C_SQ1_alias2 . `col_varchar_key` AS C_SQ1_field1 FROM ( C AS C_SQ1_alias1 INNER JOIN C AS C_SQ1_alias2 ON (C_SQ1_alias2 . `col_varchar_nokey` = C_SQ1_alias1 . `col_varchar_key` ) ) WHERE C_SQ1_alias2 . `pk` <= SQ1_alias1 . `pk` OR C_SQ1_alias2 . `col_varchar_key` != C_SQ1_alias2 . `col_varchar_nokey` ) GROUP BY SQ1_field1 , SQ1_field2  )  HAVING (field2 <= 9 OR field4 >= 6)  /* 4 */

Optimizer switch: semijoin=ON,materialization=OFF,firstmatch=OFF,loosescan=OFF

            Gliffy Diagrams

              Attachments

                Activity

                Ascending order - Click to sort in descending order
                Hide
                Permalink
                philipstoev Philip Stoev added a comment - - edited

                Re: Assertion `! is_set()' failed in Diagnostics_area::set_ok_status() with concurrent subquery workload
                To reproduce:

                 perl runall.pl --queries=100000000 --debug --mysqld1=--debug-assert-if-crashed-table --queries=100M --duration=1200 --threads=6 --seed=time --mysqld1=--log-output=file --no-mask --views --reporter=QueryTimeout,Backtrace,ErrorLog,Deadlock --mysqld1=--optimizer_switch=semijoin=ON,materialization=OFF,firstmatch=OFF,loosescan=OFF --validator=Transformer --transformers=DisableChosenPlan,ConvertLiteralsToSubqueries,ExecuteAsInsertSelect,ExecuteAsSelectItem,ExecuteAsUpdateDelete,ExecuteAsWhereSubquery,InlineSubqueries,ExecuteAsPreparedTwice --notnull --engine=MyISAM --skip-gendata --seed=1316018912  --basedir1=/home/philips/bzr/maria-5.3  --mysqld1=--init-file=`pwd`/conf/optimizer/world.sql  --grammar=bug850647.yy
                Show
                philipstoev Philip Stoev added a comment - - edited Re: Assertion `! is_set()' failed in Diagnostics_area::set_ok_status() with concurrent subquery workload To reproduce: perl runall.pl --queries=100000000 --debug --mysqld1=--debug-assert-if-crashed-table --queries=100M --duration=1200 --threads=6 --seed=time --mysqld1=--log-output=file --no-mask --views --reporter=QueryTimeout,Backtrace,ErrorLog,Deadlock --mysqld1=--optimizer_switch=semijoin=ON,materialization=OFF,firstmatch=OFF,loosescan=OFF --validator=Transformer --transformers=DisableChosenPlan,ConvertLiteralsToSubqueries,ExecuteAsInsertSelect,ExecuteAsSelectItem,ExecuteAsUpdateDelete,ExecuteAsWhereSubquery,InlineSubqueries,ExecuteAsPreparedTwice --notnull --engine=MyISAM --skip-gendata --seed=1316018912 --basedir1=/home/philips/bzr/maria-5.3 --mysqld1=--init-file=`pwd`/conf/optimizer/world.sql --grammar=bug850647.yy
                Hide
                Permalink
                philipstoev Philip Stoev added a comment - - edited

                To reproduce:

                 perl runall.pl --queries=100000000 --debug --mysqld1=--debug-assert-if-crashed-table --queries=100M --duration=1200 --threads=6 --seed=time --mysqld1=--log-output=file --no-mask --views --reporter=QueryTimeout,Backtrace,ErrorLog,Deadlock --mysqld1=--optimizer_switch=semijoin=ON,materialization=OFF,firstmatch=OFF,loosescan=OFF --validator=Transformer --transformers=DisableChosenPlan,ConvertLiteralsToSubqueries,ExecuteAsInsertSelect,ExecuteAsSelectItem,ExecuteAsUpdateDelete,ExecuteAsWhereSubquery,InlineSubqueries,ExecuteAsPreparedTwice --notnull --engine=MyISAM --skip-gendata --seed=1316018912  --basedir1=/home/philips/bzr/maria-5.3  --mysqld1=--init-file=`pwd`/conf/optimizer/world.sql  --grammar=bug850647.yy

                grammar file
                LPexportBug850647_bug850647.yy

                Show
                philipstoev Philip Stoev added a comment - - edited To reproduce: perl runall.pl --queries=100000000 --debug --mysqld1=--debug-assert-if-crashed-table --queries=100M --duration=1200 --threads=6 --seed=time --mysqld1=--log-output=file --no-mask --views --reporter=QueryTimeout,Backtrace,ErrorLog,Deadlock --mysqld1=--optimizer_switch=semijoin=ON,materialization=OFF,firstmatch=OFF,loosescan=OFF --validator=Transformer --transformers=DisableChosenPlan,ConvertLiteralsToSubqueries,ExecuteAsInsertSelect,ExecuteAsSelectItem,ExecuteAsUpdateDelete,ExecuteAsWhereSubquery,InlineSubqueries,ExecuteAsPreparedTwice --notnull --engine=MyISAM --skip-gendata --seed=1316018912 --basedir1=/home/philips/bzr/maria-5.3 --mysqld1=--init-file=`pwd`/conf/optimizer/world.sql --grammar=bug850647.yy grammar file LPexportBug850647_bug850647.yy
                Hide
                Permalink
                ratzpo Rasmus Johansson added a comment -

                Launchpad bug id: 850647

                Show
                ratzpo Rasmus Johansson added a comment - Launchpad bug id: 850647
                Hide
                Permalink
                elenst Elena Stepanova added a comment -

                There are no revision IDs for the server and RQG in the bug report, so they can only be guessed approximately by the date of reporting.
                MariaDB 5.3.1 which was released a few days before the report, and several revisions after it, fail on the same test with https://bugs.launchpad.net/maria/+bug/849717 , which was fixed by revno 3188; so, apparently to reproduce this bug, either another tree or some other conditions are needed.
                Alternatively, it's possible that the race condition itself happened when bug#849717 was hit, only the server managed to assert on create table before it crashed in another thread.

                Show
                elenst Elena Stepanova added a comment - There are no revision IDs for the server and RQG in the bug report, so they can only be guessed approximately by the date of reporting. MariaDB 5.3.1 which was released a few days before the report, and several revisions after it, fail on the same test with https://bugs.launchpad.net/maria/+bug/849717 , which was fixed by revno 3188; so, apparently to reproduce this bug, either another tree or some other conditions are needed. Alternatively, it's possible that the race condition itself happened when bug#849717 was hit, only the server managed to assert on create table before it crashed in another thread.

                  People

                  • Assignee:
                    elenst Elena Stepanova
                    Reporter:
                    philipstoev Philip Stoev
                  • Votes:
                    0 Vote for this issue
                    Watchers:
                    2 Start watching this issue

                    Dates

                    • Created:
                      Updated:
                      Resolved: