March 2015 - developers - lists.mariadb.org

[Maria-developers] Using THD instance in a daemon thread.
by 김경남 31 Mar '15

31 Mar '15

Hi, I'm writing mariadb daemon type plugin. I hava a question. In a daemon thread, I want to use THD instance. for example, ------------------------------------------------------------------------ ----------- class THD; static int plugin_init() { ... pthread_create(&id,&attr,my_func,...); } void my_func(void* p) { THD* thd = new THD; thd->thread_stack= (char*) &thd; thd->store_globals(); tables.init_one_table("performance_schema", 18, "events_statements_current", 25, "events_statements_current", TL_READ); open_and_lock_tables(thd, &tables, FALSE, MYSQL_LOCK_IGNORE_TIMEOUT); ... } ------------------------------------------------------------------------ -------------------- but In my_func(), new THD cause error. I guess thread problem. So I moved that code to plugin_init() for test as below -------------------------------------------------------------- class THD; static int plugin_init() { ... pthread_create(&id,&attr,my_func,...); THD* thd = new THD; thd->thread_stack= (char*) &thd; thd->store_globals(); tables.init_one_table("performance_schema", 18, "events_statements_current", 25, "events_statements_current", TL_READ); open_and_lock_tables(thd, &tables, FALSE, MYSQL_LOCK_IGNORE_TIMEOUT); ... } ------------------------------------------------------------------------ ----- and then there is no problem. but I need a daemon thread. I think that the reason of this error is assertion of thread. So I changed my code as below. ------------------------------------------------------------------------ --------- class THD; THD* thd; static int plugin_init() { ... pthread_create(&id,&attr,my_func,...); THD* thd = new THD; // moved here from my_func() for avoiding error. } void my_func(void* p) { thd->thread_stack= (char*) &thd; thd->store_globals(); tables.init_one_table("performance_schema", 18, "events_statements_current", 25, "events_statements_current", TL_READ); open_and_lock_tables(thd, &tables, FALSE, MYSQL_LOCK_IGNORE_TIMEOUT); ... } ------------------------------------------------------------------------ ----------------- At this time, thd->store_globals() cause error. In sql_class.cc , mysys_var->id= thread_id cause error because mysys_var is null. Let me know how to use THD instance in daemon thread. Regards, Kyungnam 김경남 부책임연구원 솔루션사업본부 정보기술 연구원 신기술개발팀 02-2099-6177 / 010-2208-3006 _____ 상기 메시지와 첨부화일 내에는 비밀정보가 포함되어 있을 수 있으며, 지정된 수신자에 한하여 조회 및 사용될 수 있습니다. 만약 송신자의 실수로 인하여 상기 메시지를 수신하였다면, 송신자에게 메시지를 반송해 주시고, 원본 메시지와 모든 사본을 폐기해 주시기 바랍니다. 상기 메시지의 전체 또는 일부에 대해 무단 열람, 사용, 공개, 배포하는 것은 금지되어 있습니다.(주)LG CNS. This message and its attachments may contain confidential information, and they are intended to be viewed or used by only the individuals specified in the message. If you have received this message in an error from the sender, please contact the sender immediately to notify the error and delete all of the message and its copies. It is prohibited to view, use, make public and/or distribute part or whole of this message without written permission.

2 1

[Maria-developers] ok to change FixVersion to 10.1 for MDEV-7649 wrong result when comparing utf8 column with an invalid literal
by Alexander Barkov 31 Mar '15

31 Mar '15

Hi Sergei, Do you think it's ok to change FixVersion for MDEV-7649 to 10.1? It requires some changes in sql_lex.cc, sql_yacc.yy and needs a new Item_string-alike Item to handle such malformed strings properly. Too big change to 10.0 or 5.5.

2 1

[Maria-developers] TokuDB test suite failures when building at Launchpad
by Otto Kekäläinen 29 Mar '15

29 Mar '15

Hello! Are here any TokuDB devs who could help me debug why TokuDB does repeatedly fails to pass build suite tests when built on Launchpad.net? I repeatedly get these test failures when running MariaDB builds on Launchpad: rpl-tokudb.tokudb_innodb_xa_crash tokudb_bugs.mdev4533 tokudb.simple_join_tokudb_innodb tokudb.truncate_txn_rollback_innodb tokudb_bugs.5733_innodb The build run log can be viewed at https://launchpad.net/~mysql-ubuntu/+archive/ubuntu/mariadb-10.0/+builds?bu… and it links to the individual build logs. All amd64 builds fail while all i386 (where TokuDB is disabled) pass OK. Sounds very much like a TokuDB-related issue. Or could the problem be something else and the TokuDB tests are just the first victim? Example of a recent log: https://launchpadlibrarian.net/196425361/buildlog_ubuntu-vivid-amd64.mariad… One example of the test outputs: *** rpl-tokudb.tokudb_innodb_xa_crash 'mix,xtradb' [ fail ] Test ended at 2015-02-02 22:29:06 CURRENT_TEST: rpl-tokudb.tokudb_innodb_xa_crash Failed to start mysqld.1 mysqltest failed but provided no output - saving '/build/buildd/mariadb-10.0-10.0.16/builddir/mysql-test/var/log/rpl-tokudb.tokudb_innodb_xa_crash-mix,xtradb/' to '/build/buildd/mariadb-10.0-10.0.16/builddir/mysql-test/var/log/rpl-tokudb.tokudb_innodb_xa_crash-mix,xtradb/' - found 'core' (1/5) Trying 'dbx' to get a backtrace Trying 'gdb' to get a backtrace Compressed file /build/buildd/mariadb-10.0-10.0.16/builddir/mysql-test/var/log/rpl-tokudb.tokudb_innodb_xa_crash-mix,xtradb/mysqld.1/data/core ***Warnings generated in error logs during shutdown after running tests: rpl-tokudb.tokudb_innodb_xa_crash 150202 22:29:05 [ERROR] mysqld got signal 11 ; Attempting backtrace. You can use the following information to find out 150202 22:29:05 [ERROR] mysqld got signal 11 ; Attempting backtrace. You can use the following information to find out ***

4 15

[Maria-developers] GSoC Proposal Submitted.
by arlene seuleu 28 Mar '15

28 Mar '15

Hello everyone, I am Arlene Flavie, currently a Sofeware Engineering student and I have submitted a proposal for the CONC-125 task: "Imporrt and Export popular data formats from and to dynamic columns". I had been in contact with the mentor and other mariadb developpers(serg, danblack) whose help was immense in writing my proposal. However I am still available to discuss it here and on #maria as Arlene. Hoping to get feedback from the community soon. Best Regards Arlene Flavie

1 0

[Maria-developers] MySQL 5.7.6 adds "statement digest" independent of P_S
by Peter Laursen 27 Mar '15

27 Mar '15

.. as well as configurable digest length. http://mysqlserverteam.com/mysql-statement-digests-configurable-max-lengths/ http://mysqlserverteam.com/mysql-statement-digests/ Will this be ported to MariaDB? IMO this is/was one of the (not so many) useful things in P_S. And better now. -- Peter -- Webyog

1 0

[Maria-developers] In daemon plugin, THD problem.
by 김경남 27 Mar '15

27 Mar '15

Hi, Sergei! I'm writing daemon plugin . I hava a problem and don't find soultion yet. There is my code as below. ex)plugin.cc ----------------------------------------------------- pthread_handler_t monm_monitor(void *p) { .... THD* thd; thd = new THD; thd->thread_stack= (char*) &thd; tables.init_one_table("performance_schema", 18, "events_statements_current", 25, "events_statements_current", TL_READ); open_and_lock_tables(thd, &tables, FALSE, MYSQL_LOCK_IGNORE_TIMEOUT); ..... } static int monm_plugin_init(void *p __attribute__ ((unused))) { ... pthread_create(&mon_thread, &attr, monm_monitor,(void *) ctx)!= 0) .... } ------------------------------------------------------------- In monm_monitor function, THD* thd cause server crash. so I examined coredump. #6 0x0000000000598752 in THD::THD (this=0x7f9b64000930, is_wsrep_applier=false) at /home/windfree/workspace/ws-mariadb/maria-mon/sql/sql_class.cc:1015 I don't know the reason of crash. What I just want to is read events_statements_current's row by using internal interface in a daemon plugin. Is there a another way ? let me know the solution. Regards, Kyungnam. 김경남 부책임연구원 솔루션사업본부 정보기술 연구원 신기술개발팀 02-2099-6177 / 010-2208-3006 _____ 상기 메시지와 첨부화일 내에는 비밀정보가 포함되어 있을 수 있으며, 지정된 수신자에 한하여 조회 및 사용될 수 있습니다. 만약 송신자의 실수로 인하여 상기 메시지를 수신하였다면, 송신자에게 메시지를 반송해 주시고, 원본 메시지와 모든 사본을 폐기해 주시기 바랍니다. 상기 메시지의 전체 또는 일부에 대해 무단 열람, 사용, 공개, 배포하는 것은 금지되어 있습니다.(주)LG CNS. This message and its attachments may contain confidential information, and they are intended to be viewed or used by only the individuals specified in the message. If you have received this message in an error from the sender, please contact the sender immediately to notify the error and delete all of the message and its copies. It is prohibited to view, use, make public and/or distribute part or whole of this message without written permission.

2 2

Re: [Maria-developers] [Org FW][RE]Re: init_one_table problem.
by sergei 25 Mar '15

25 Mar '15

Hi, 김경남! On Mar 25, 김경남 wrote: > Hi, Sergei ! > > below contents is my previous question . > > After sending email , I see the debug message.like below. > > "Assertion thd == current_thd() failed." > > My code contains. > > THD thd = new THD; > > so I changed my code as > > THD thd = current_thd; > > but thd is null. > > how to I fix it? Note that sql_acl.cc (for example) has THD *thd=new THD; thd->thread_stack= (char*) &thd; thd->store_globals(); All three lines are important. In particular, THD::store_global sets current_thd. Regards, Sergei

1 0

[Maria-developers] init_one_table problem.
by 김경남 25 Mar '15

25 Mar '15

Hi ~ I hava a problem below code in my plugin. ---------------------------------- TABLE_LIST tables; TABLE table; tables.init_one_table(C_STRING_WITH_LEN("performance_schema"),C_STRING_W ITH_LEN("events_statements_current"), events_statements_current, TL_READ); if(table = tables.table) {...} ------------------------------------------------------- During debugging, I found that tables.table is null. Is there a additional task before my code? Regards, Kyungnam. 김경남 부책임연구원 솔루션사업본부 정보기술 연구원 신기술개발팀 02-2099-6177 / 010-2208-3006 _____ 상기 메시지와 첨부화일 내에는 비밀정보가 포함되어 있을 수 있으며, 지정된 수신자에 한하여 조회 및 사용될 수 있습니다. 만약 송신자의 실수로 인하여 상기 메시지를 수신하였다면, 송신자에게 메시지를 반송해 주시고, 원본 메시지와 모든 사본을 폐기해 주시기 바랍니다. 상기 메시지의 전체 또는 일부에 대해 무단 열람, 사용, 공개, 배포하는 것은 금지되어 있습니다.(주)LG CNS. This message and its attachments may contain confidential information, and they are intended to be viewed or used by only the individuals specified in the message. If you have received this message in an error from the sender, please contact the sender immediately to notify the error and delete all of the message and its copies. It is prohibited to view, use, make public and/or distribute part or whole of this message without written permission.

2 1

Re: [Maria-developers] [RE]Re: I wonder what elements add to CMakeLists.txt
by Sergei Golubchik 25 Mar '15

25 Mar '15

Hi, kyungnam! On Mar 24, kyungnam wrote: > > I hava a question. > > I'm developing mariadb plugin . and I make a development environment > according to below url. > > https://mariadb.com/blog/compiling-debugging-mariadband-mysql-eclipse-scrat… > > I seperate source folder and package folder. > > During developing, I hava a trouble with too much build time. > > for example, after I edit sql_parse.cc (add one or two lines) , I run > command "make" under source root folder. > > The build time takes about 5 ~ 10min. and then I run command "make > package". That time takes about 10min. > > At first, I copied libmysqld.so.18 to package folder. but had no > effect. finally I run command "make package" and reinstall mariadb. > > I think I'm doing wrong things. > > please let me know the efficient way to build and debug. I'm not using Eclipse, so I'm afraid I cannot help you with that. If you just modify sql_parse.cc, the compilation should be fast and only that file has to be recompiled (and few binaries re-linked). If you see a lot of files recompiled than either you modified more than sql_parse.cc or your IDE is doing that for you. Make package needs to copy many files so it's a bit slow if you do it very often. I don't - unless I specifically work on packaging I run "make package" very rarely. You don't have to compile libmysqld every time. Run cmake -DWITH_EMBEDDED_SERVER=OFF in your build directory. Regards, Sergei

1 0

[Maria-developers] Merge 10.0 -> 10.1
by Kristian Nielsen 25 Mar '15

25 Mar '15

I have some patches in 10.0 that have not yet been merged to 10.1. Today I attempted the merge, but there were a few conflicts that I was not sure how to handle. Some charset-related functions touching fields that had moved to another struct, some error handling functions in XtraDB with new extra arguments. In case someone else does the merge, I just wanted to point out one conflict in one of my patches in the 10.0 -> 10.1 merge, as the correct resolution is not obvious: <<<<<<< HEAD if (mysql_bin_log.is_open()) return true; ||||||| merged common ancestors ======= if (mysql_bin_log.is_open() && opt_binlog_commit_wait_count > 0) return true; >>>>>>> origin/10.0 The correct conflict resolution here is to leave the code as it was in 10.1: if (mysql_bin_log.is_open()) return true; Or I can do the complete 10.0 -> 10.1 merge if necessary, figuring out the other conflicts, but I did not want to do that without coordinating, to avoid duplicating efforts. - Kristian.

2 1