- developers - lists.mariadb.org

Re: [Maria-developers] [Commits] b904ca0: MDEV-9297 - mysql_install_db launches mysqld 3x
by Sergei Golubchik 29 Dec '15

29 Dec '15

Hi, Sergey! On Dec 23, Sergey Vojtovich wrote: > revision-id: b904ca0c2e9e35ffecb314f1082e45bfb42038f3 (mariadb-10.1.8-99-gb904ca0) > parent(s): 29290a3d53a05bd7177bcf7de05e1c8fff6e9cea > committer: Sergey Vojtovich > timestamp: 2015-12-23 15:52:34 +0400 > message: > > MDEV-9297 - mysql_install_db launches mysqld 3x > > Perform database installation in one mysqld run. This makes mysql_install_db > 3x faster. > > diff --git a/scripts/mysql_install_db.sh b/scripts/mysql_install_db.sh > index f8f36bf..9dcd23a 100644 > --- a/scripts/mysql_install_db.sh > +++ b/scripts/mysql_install_db.sh > @@ -427,7 +427,7 @@ mysqld_install_cmd_line() > > # Create the system and help tables by passing them to "mysqld --bootstrap" > s_echo "Installing MariaDB/MySQL system tables in '$ldata' ..." > -if { echo "use mysql;"; cat "$create_system_tables" "$create_system_tables2" "$fill_system_tables"; } | eval "$filter_cmd_line" | mysqld_install_cmd_line > /dev/null > +if { echo "use mysql;"; cat "$create_system_tables" "$create_system_tables2" "$fill_system_tables" "$fill_help_tables" "$maria_add_gis_sp"; } | eval "$filter_cmd_line" | mysqld_install_cmd_line > /dev/null > then > s_echo "OK" > else > @@ -462,26 +462,6 @@ else > exit 1 > fi > > -s_echo "Filling help tables..." > -if { echo "use mysql;"; cat "$fill_help_tables"; } | mysqld_install_cmd_line > /dev/null > -then > - s_echo "OK" > -else > - echo > - echo "WARNING: HELP FILES ARE NOT COMPLETELY INSTALLED!" > - echo "The \"HELP\" command might not work properly." > -fi > - > -s_echo "Creating OpenGIS required SP-s..." > -if { echo "use test;"; cat "$maria_add_gis_sp"; } | mysqld_install_cmd_line > /dev/null > -then > - s_echo "OK" > -else > - echo > - echo "WARNING: OPENGIS REQUIRED SP-S WERE NOT COMPLETELY INSTALLED!" > - echo "GIS extentions might not work properly." > -fi Uhm, I don't know. I believe the point was not to treat fill_help_tables failure as fatal. To keep this logic you can merge maria_add_gis_sp and fill_help_tables in one mysqld_install_cmd_line invocation. But that's all. On the other hand, perhaps it's not important to preserve this logic? Then your patch is fine... Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org -- Vote for my Percona Live 2016 talks: https://www.percona.com/live/data-performance-conference-2016/sessions/mari… https://www.percona.com/live/data-performance-conference-2016/sessions/mari…

2 2

Re: [Maria-developers] [Commits] 05c122b: MDEV-8491 - On shutdown, report the user and the host executed that.
by Sergei Golubchik 28 Dec '15

28 Dec '15

Hi, Sergey! On Dec 24, Sergey Vojtovich wrote: > revision-id: 05c122b3ccd0102a3414a7d59313e6662a1fd5e4 (mariadb-10.1.8-101-g05c122b) > parent(s): 7c533a17523d2d6ef380487ab35151da0208090f > committer: Sergey Vojtovich > timestamp: 2015-12-24 19:53:44 +0400 > message: > > MDEV-8491 - On shutdown, report the user and the host executed that. ok to push Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org -- Vote for my Percona Live 2016 talks: https://www.percona.com/live/data-performance-conference-2016/sessions/mari… https://www.percona.com/live/data-performance-conference-2016/sessions/mari…

1 0

[Maria-developers] Option to _not_ disable semisync on slave timeout?
by Kristian Nielsen 28 Dec '15

28 Dec '15

Hi Jonas (or anyone else who may know), If I understand correctly, semisync replication has a timeout for how long it will wait for a slave to acknowledge transactions. If the timeout is exceeded, semisync is turned off on the master. Do you know if there is a facility to avoid semisync being turned off automatically in this case, maybe instead fail and roll back the master transaction? I was thinking at Google you might already have needed and implemented this (merged or not merged into MariaDB yet), or you may be aware of a patch by someone else for this? Thanks, and merry Christmas! - Kristian.

3 2

Re: [Maria-developers] [Commits] a1ab431: MDEV-7261 - Backport the default autosized value of
by Sergei Golubchik 28 Dec '15

28 Dec '15

Hi, Sergey! On Nov 20, Sergey Vojtovich wrote: > revision-id: a1ab4314d1f88fa954a774c322709822d7b95344 (mariadb-10.1.8-58-ga1ab431) > parent(s): cfc135b645db5fa78b9f63b1959ea3f2be137a26 > committer: Sergey Vojtovich > timestamp: 2015-11-20 17:32:26 +0400 > message: > > MDEV-7261 - Backport the default autosized value of > "table_definition_cache" from MySQL 5.6.8+ > > Fixed table_definition_cache auto-sizing. Note that this patch doesn't > address broken PFS auto-sizing. Do you plan to address it after you've broken it? How? Regards, Sergei

2 9

[Maria-developers] ABI breakage in Connector/C sometime between 2.0.0 and 2.2.2
by Daniel Schepler 27 Dec '15

27 Dec '15

While preparing Debian packaging for MariaDB Connector/C version 2.2.2, I noticed that there's an ABI breakage without a corresponding change of SONAME: version 2.0.0 had a symbol "get_tty_password@Base" whereas version 2.2.2 doesn't. I'm wondering how big of a breakage it is - i.e. is that something that was never meant to be in the API/ABI in the first place, or is an oversight? I'm also wondering whether it's worth renaming the Debian package from libmariadb2 to libmariadb2a because of this. -- Daniel Schepler

1 0

Re: [Maria-developers] [Commits] 2b3bffb: MDEV-8491 - On shutdown, report the user and the host executed that.
by Sergei Golubchik 24 Dec '15

24 Dec '15

Hi, Sergey! On Nov 25, Sergey Vojtovich wrote: > revision-id: 2b3bffb42cc4adbec2d91c1b9c4028374b63a51a (mariadb-10.1.8-75-g2b3bffb) > parent(s): 6019fee7d84e8ec7d64337ad080a04f9c106bb68 > committer: Sergey Vojtovich > timestamp: 2015-11-25 18:12:19 +0400 > message: > > MDEV-8491 - On shutdown, report the user and the host executed that. First, a test would be nice. I suppose you can add it to the main.shutdown test. > diff --git a/sql/mysqld.cc b/sql/mysqld.cc > index 0502841..0694cb0 100644 > --- a/sql/mysqld.cc > +++ b/sql/mysqld.cc > @@ -1806,10 +1806,34 @@ static void close_server_sock() > #endif /*EMBEDDED_LIBRARY*/ > > > -void kill_mysql(void) > +static volatile char *shutdown_user; > +static void set_shutdown_user(THD *thd) > +{ > + char user_host_buff[MAX_USER_HOST_SIZE + 1]; > + char *user, *expected_shutdown_user= 0; > + Security_context *sctx= thd->security_ctx; > + > + strxnmov(user_host_buff, MAX_USER_HOST_SIZE, > + sctx->priv_user ? sctx->priv_user : "", "[", > + sctx->user ? sctx->user : (thd->slave_thread ? "SQL_SLAVE" : ""), > + "] @ ", > + sctx->host ? sctx->host : "", " [", > + sctx->ip ? sctx->ip : "", "]", NullS); this will be the fifth time this expression shows up in the source code. turn it into a function, perhaps? Like, an inline method of THD? also, sometimes it's "SQL_SLAVE", sometimes it's "", sometimes it's "unauthenticated", etc. I hope these differences can be removed and there will be one method that works identically for all occasions. Btw, you can use safe_str() helper here. > + > + if ((user= my_strdup(user_host_buff, MYF(0))) && > + !my_atomic_casptr((void **) &shutdown_user, > + (void **) &expected_shutdown_user, user)) > + my_free(user); > +} Interesting. Why is that? > + > + > +void kill_mysql(THD *thd) > { > DBUG_ENTER("kill_mysql"); > > + if (thd) > + set_shutdown_user(thd); > + > #if defined(SIGNALS_DONT_BREAK_READ) && !defined(EMBEDDED_LIBRARY) > abort_loop=1; // Break connection loops > close_server_sock(); // Force accept to wake up > @@ -1888,7 +1912,13 @@ static void __cdecl kill_server(int sig_ptr) > if (sig != 0) // 0 is not a valid signal number > my_sigset(sig, SIG_IGN); /* purify inspected */ > if (sig == MYSQL_KILL_SIGNAL || sig == 0) > - sql_print_information(ER_DEFAULT(ER_NORMAL_SHUTDOWN),my_progname); > + { > + char *user= (char *) my_atomic_loadptr((void**) &shutdown_user); > + sql_print_information(ER_DEFAULT(ER_NORMAL_SHUTDOWN), my_progname, > + user ? user : "unknown"); can it, really, be "unknown" here? when? > + if (user) > + my_free(user); > + } > else > sql_print_error(ER_DEFAULT(ER_GOT_SIGNAL),my_progname,sig); /* purecov: inspected */ > > diff --git a/sql/share/errmsg-utf8.txt b/sql/share/errmsg-utf8.txt > index 59908dc..f65d466 100644 > --- a/sql/share/errmsg-utf8.txt > +++ b/sql/share/errmsg-utf8.txt > @@ -1735,29 +1735,29 @@ ER_WRONG_AUTO_KEY 42000 S1009 > ER_UNUSED_9 > eng "You should never see it" > ER_NORMAL_SHUTDOWN > - cze "%s: normální ukončení\n" > - dan "%s: Normal nedlukning\n" > - nla "%s: Normaal afgesloten \n" > - eng "%s: Normal shutdown\n" > - est "%s: MariaDB lõpetas\n" > - fre "%s: Arrêt normal du serveur\n" > - ger "%s: Normal heruntergefahren\n" > - greek "%s: Φυσιολογική διαδικασία shutdown\n" > - hun "%s: Normal leallitas\n" > - ita "%s: Shutdown normale\n" > - jpn "%s: 通常シャットダウン\n" > - kor "%s: 정상적인 shutdown\n" > - nor "%s: Normal avslutning\n" > - norwegian-ny "%s: Normal nedkopling\n" > - pol "%s: Standardowe zakończenie działania\n" > - por "%s: 'Shutdown' normal\n" > - rum "%s: Terminare normala\n" > - rus "%s: Корректная остановка\n" > - serbian "%s: Normalno gašenje\n" > - slo "%s: normálne ukončenie\n" > - spa "%s: Apagado normal\n" > - swe "%s: Normal avslutning\n" > - ukr "%s: Нормальне завершення\n" > + cze "%s: normální ukončení %s\n" > + dan "%s: Normal nedlukning %s\n" > + nla "%s: Normaal afgesloten %s\n" > + eng "%s: Normal shutdown by %s\n" > + est "%s: MariaDB lõpetas %s\n" > + fre "%s: Arrêt normal du serveur %s\n" > + ger "%s: Normal heruntergefahren %s\n" > + greek "%s: Φυσιολογική διαδικασία shutdown %s\n" > + hun "%s: Normal leallitas %s\n" > + ita "%s: Shutdown normale %s\n" > + jpn "%s: 通常シャットダウン %s\n" > + kor "%s: 정상적인 shutdown %s\n" > + nor "%s: Normal avslutning %s\n" > + norwegian-ny "%s: Normal nedkopling %s\n" > + pol "%s: Standardowe zakończenie działania %s\n" > + por "%s: 'Shutdown' normal %s\n" > + rum "%s: Terminare normala %s\n" > + rus "%s: Корректная остановка пользователем %s\n" > + serbian "%s: Normalno gašenje %s\n" > + slo "%s: normálne ukončenie %s\n" > + spa "%s: Apagado normal %s\n" > + swe "%s: Normal avslutning %s\n" > + ukr "%s: Нормальне завершення %s\n" Here you've fixed only the english and russian messages. Others would look pretty weird. Normal procedure in such a case is to remove all other languages and only keep those that you've fixed. In this case you can, perhaps, fix all of them in some easy way, like ger "%s: Normal heruntergefahren (%s)\n" ger "%s: Normal heruntergefahren / %s\n" ger "%s (%s): Normal heruntergefahren\n" ger "%s / %s: Normal heruntergefahren\n" or something else along these lines. See how it'll look like in the log: mysqld: Normal heruntergefahren foo [bar] @ localhost [127.0.0.1] mysqld: Normal heruntergefahren (foo [bar] @ localhost [127.0.0.1]) mysqld: Normal heruntergefahren / foo [bar] @ localhost [127.0.0.1] mysqld (foo [bar] @ localhost [127.0.0.1]): Normal heruntergefahren mysqld / foo [bar] @ localhost [127.0.0.1]: Normal heruntergefahren Regards, Sergei

2 3

Re: [Maria-developers] [Commits] ed5ec30: MDEV-7780 - Support for faking server version
by Sergei Golubchik 24 Dec '15

24 Dec '15

Hi, Sergey! On Dec 24, Sergey Vojtovich wrote: > revision-id: ed5ec30f9ca448086db5acbfd5f484a534119810 (mariadb-10.1.8-101-ged5ec30) > parent(s): 1f19c6aa93de730cd10c9fcbb337f9ec5c30ef9d > committer: Sergey Vojtovich > timestamp: 2015-12-24 15:46:18 +0400 > message: > > MDEV-7780 - Support for faking server version > > Embedded now supports "--version=<version>", while "--version" is still silently > ignored. Also only run protocol check in non-embedded mode. As I said on irc, I'd simply use not_embedded.inc, but ok, if you want it to work in embedded, sure, do it. One question below. > --- > mysql-test/suite/sys_vars/t/version.test | 3 ++- > sql/mysqld.cc | 4 ++-- > 2 files changed, 4 insertions(+), 3 deletions(-) > > diff --git a/mysql-test/suite/sys_vars/t/version.test b/mysql-test/suite/sys_vars/t/version.test > index 7955bf8..daa9538 100644 > --- a/mysql-test/suite/sys_vars/t/version.test > +++ b/mysql-test/suite/sys_vars/t/version.test > @@ -1,5 +1,6 @@ > SELECT @@version; > perl; > - $cnt= grep /my_favorite_version/, `$ENV{MYSQL} -e status`; > + $cnt= $ENV{MYSQL_TEST} =~ /mysqltest_embedded / ? 1 : > + grep /my_favorite_version/, `$ENV{MYSQL} -e status`; Why wouldn't you use mysql_embedded in this case? Like $mysql= $ENV{MYSQL_TEST} =~ /mysqltest_embedded / ? 'MYSQL_EMBEDDED' : 'MYSQL'; $cnt= grep /my_favorite_version/, `$ENV{$mysql} -e status`; > print "$cnt\n"; > EOF > diff --git a/sql/mysqld.cc b/sql/mysqld.cc > index 7360204..fa3a70f 100644 > --- a/sql/mysqld.cc > +++ b/sql/mysqld.cc > @@ -8972,20 +8972,20 @@ mysqld_get_one_option(int optid, const struct my_option *opt, char *argument) > binlog_format_used= true; > break; > #include <sslopt-case.h> > -#ifndef EMBEDDED_LIBRARY > case 'V': > if (argument) > { > strmake(server_version, argument, sizeof(server_version) - 1); > set_sys_var_value_origin(&server_version_ptr, sys_var::CONFIG); > } > +#ifndef EMBEDDED_LIBRARY > else > { > print_version(); > opt_abort= 1; // Abort after parsing all options > } > - break; > #endif /*EMBEDDED_LIBRARY*/ > + break; > case 'W': > if (!argument) > global_system_variables.log_warnings++; Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org -- Vote for my Percona Live 2016 talks: https://www.percona.com/live/data-performance-conference-2016/sessions/mari… https://www.percona.com/live/data-performance-conference-2016/sessions/mari…

2 2

Re: [Maria-developers] [Commits] 370ab48: MDEV-9209 - [PATCH] scripts: Do not prepend the prefix to absolute paths
by Sergei Golubchik 22 Dec '15

22 Dec '15

Hi, Sergey! On Dec 18, Sergey Vojtovich wrote: > revision-id: 370ab48e9fcf4bc15ffa5e313fc34aa97981a832 (mariadb-10.1.9-24-g370ab48) > parent(s): 428e09a789a17211de9b4f0a2c4c1226d5dcf993 > committer: Sergey Vojtovich > timestamp: 2015-12-18 16:50:03 +0400 > message: > > MDEV-9209 - [PATCH] scripts: Do not prepend the prefix to absolute paths > > Allow absolute paths for INSTALL_*DIR. Okay, I agree now. Let's use your approach :) See below, one minor issue in mariadb.pc file. > diff --git a/scripts/CMakeLists.txt b/scripts/CMakeLists.txt > index eb78016..54289c0 100644 > --- a/scripts/CMakeLists.txt > +++ b/scripts/CMakeLists.txt > > IF(UNIX) > CONFIGURE_FILE(${CMAKE_CURRENT_SOURCE_DIR}/mysql_install_db.sh > ${CMAKE_CURRENT_BINARY_DIR}/mysql_install_db ESCAPE_QUOTES @ONLY) > - SET(DEST ${INSTALL_SCRIPTDIR}) > SET(EXT) you can remove EXT as well, it's apparently not used anymore. > EXECUTE_PROCESS( > COMMAND chmod +x ${CMAKE_CURRENT_BINARY_DIR}/mysql_install_db > diff --git a/support-files/mariadb.pc.in b/support-files/mariadb.pc.in > index bc84a5c..0605d86 100644 > --- a/support-files/mariadb.pc.in > +++ b/support-files/mariadb.pc.in > @@ -1,18 +1,18 @@ > # these four variables are present in almost every .pc file > prefix=@CMAKE_INSTALL_PREFIX@ > exec_prefix=${prefix} > -libdir=${prefix}/@INSTALL_LIBDIR@ > -includedir=${prefix}/@INSTALL_INCLUDEDIR@ > +libdir=@INSTALL_LIBDIRABS@ > +includedir=@INSTALL_INCLUDEDIRABS@ This is, strictly speaking, incorrect. In the old version one can change prefix to any other value and all other paths will auto-adjust. See: $ pkg-config --libs openrc -L/lib64 -lrc $ pkg-config --libs openrc --define-variable=prefix=/opt -L/opt/lib64 -lrc Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org -- Vote for my Percona Live 2016 talks: https://www.percona.com/live/data-performance-conference-2016/sessions/mari… https://www.percona.com/live/data-performance-conference-2016/sessions/mari…

2 2

Re: [Maria-developers] [Commits] daad482: MDEV-8615: Assertion `m_cpp_buf <= begin_ptr && begin_ptr <= m_cpp_buf + m_buf_length' failed in Lex_input_stream::body_utf8_start
by Sergei Golubchik 22 Dec '15

22 Dec '15

Hi, Sanja! On Dec 21, OleksandrByelkin wrote: > revision-id: daad4828b078485ee7ac60af4fc165ae35684395 (mariadb-10.1.9-21-gdaad482) > parent(s): 59fcd7ff2315d007045eb987da5f21abbea6f6f1 > committer: Oleksandr Byelkin > timestamp: 2015-12-21 20:13:12 +0100 > message: > > MDEV-8615: Assertion `m_cpp_buf <= begin_ptr && begin_ptr <= m_cpp_buf + m_buf_length' failed in Lex_input_stream::body_utf8_start > > If it is EOF then it may not be start of compound statement. > > --- > mysql-test/r/compound.result | 7 +++++++ > mysql-test/t/compound.test | 7 +++++++ > sql/sql_yacc.yy | 7 ++++++- > 3 files changed, 20 insertions(+), 1 deletion(-) > > diff --git a/mysql-test/r/compound.result b/mysql-test/r/compound.result > index 92d3226..1d412e6 100644 > --- a/mysql-test/r/compound.result > +++ b/mysql-test/r/compound.result > @@ -162,3 +162,10 @@ a > begin not atomic select a from t1 having a > 1; end| > a > drop table t1| > +# > +# MDEV-8615: Assertion `m_cpp_buf <= begin_ptr && > +# begin_ptr <= m_cpp_buf + m_buf_length' failed in > +# Lex_input_stream::body_utf8_start > +# > +b'| > +ERROR 42000: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'b'' at line 1 > diff --git a/mysql-test/t/compound.test b/mysql-test/t/compound.test > index 1c5d12a..cabdf96 100644 > --- a/mysql-test/t/compound.test > +++ b/mysql-test/t/compound.test > @@ -150,3 +150,10 @@ select a from t1 having a > 1| > begin not atomic select a from t1 having a > 1; end| > drop table t1| > > +--echo # > +--echo # MDEV-8615: Assertion `m_cpp_buf <= begin_ptr && > +--echo # begin_ptr <= m_cpp_buf + m_buf_length' failed in > +--echo # Lex_input_stream::body_utf8_start > +--echo # > +--error ER_PARSE_ERROR > +--query b' > diff --git a/sql/sql_yacc.yy b/sql/sql_yacc.yy > index a5a62ae..6611411 100644 > --- a/sql/sql_yacc.yy > +++ b/sql/sql_yacc.yy > @@ -237,6 +237,11 @@ static sp_head *make_sp_head(THD *thd, sp_name *name, > > static bool maybe_start_compound_statement(THD *thd) > { > + if (YYLIP->eof()) > + { > + my_parse_error(thd, ER_SYNTAX_ERROR); > + return 1; > + } no, I don't like that. Syntax errors should be issued by the parser, they should follow from the grammar, not be hard-coded in some explicit checks in the code. The problem in the current grammar - as I see it, that a code block is executed before any keyword is matched. This code block assumes it's the sp_unlabeled_control rule, while it could be anything. The fix would be to remove this code block and put this code after the first keyword is matched. > if (!thd->lex->sphead) > { > if (!make_sp_head(thd, NULL, TYPE_ENUM_PROCEDURE)) > @@ -4807,7 +4812,7 @@ create_body: > | opt_create_table_options opt_create_partitioning opt_create_select {} > /* > the following rule is redundant, but there's a shift/reduce > - conflict that prevents the rule above from parsing a syntax like > + conflict that prevents the rule above from parsing a synax like This is a strange change :) > CREATE TABLE t1 (SELECT 1); > */ Regards, Sergei Chief Architect MariaDB and security(a)mariadb.org -- Vote for my Percona Live 2016 talks: https://www.percona.com/live/data-performance-conference-2016/sessions/mari… https://www.percona.com/live/data-performance-conference-2016/sessions/mari…

2 2

[Maria-developers] MDEV-8844 Unreadable control characters printed as is in warnings
by Alexander Barkov 22 Dec '15

22 Dec '15

Hi Sergei, Please review a patch for MDEV-8844. Thanks!

2 2