- discuss - lists.mariadb.org

[Maria-discuss] ALTER TABLE removes check constraint
by Peter Laursen 22 Nov '17

22 Nov '17

Test case: CREATE TABLE checks (a INT CHECK (a>2), b INT CHECK (b>2), CONSTRAINT a_greater CHECK (a>b)); SHOW CREATE TABLE checks; ALTER TABLE `test`.`checks` CHANGE `a` `a` INT(11) NULL; SHOW CREATE TABLE checks; /' -- and now one constraint is gone! CREATE TABLE `checks` ( `a` int(11) DEFAULT NULL, `b` int(11) DEFAULT NULL CHECK (`b` > 2), CONSTRAINT `a_greater` CHECK (`a` > `b`) ) ENGINE=InnoDB DEFAULT CHARSET=latin1 */ This docs page https://mariadb.com/kb/en/library/constraint/ does not say that (unnamed) CHECK constraints (ie. defined "column-wise") may be removed like this. In this case the ALTER changes one column to what it already was and still the CONSTRAINT is removed. A more realistic scenario is that someone has an INT column that is about to overflow and wants to change it to a BIGINT. I am not sure if this is an oversight/a bug or not. But if it is not, I think the docs page should clearly state that "an unnamed CHECK constraint is droppped implicitly when an ALTER TABLE statement is executed and this ALTER TABLE statement references the particular column and the ALTER TABLE statement does not recreate the CONSTRAINT". There will be other cases where such ALTER TABLE would violate the CONSTRAINT due to already existing data. Wasn't it better if the ALTER TABLE failed then, rather than dropping the CONSTRAINT? What do standards say? How do other RDBMS behave with this? Does somebody know more about this? What say? -- Peter -- Webyog

3 5

[Maria-discuss] innodb and corrupted tables
by Federico Razzoli 20 Nov '17

20 Nov '17

I just noticed this old MySQL bug:https://bugs.mysql.com/bug.php?id=10132 In the comments (2005 to 2014) everyone seems to agree that current behaviour should be treated as a bug.The main commenter now works for MariaDB.What is currently your opinion? Will this behaviour change at some point? If so, what will be the consequences for Galera? Federico

3 3

[Maria-discuss] MariaDB GUI: BSL 1.1 - Debate
by Juan Telleria 19 Nov '17

19 Nov '17

Dear MariaDb Developers, I am writing because I would like to open a debate whether if MariaDB, now that has gained investment from Microsoft, Alibaba, and the European Union, shall create its own SQL Graphical User Interface, as good as (or better than) MySQL Workbench. I agree that MariaDB includes by default HeidiSQL in its Windows builds, but in my opinion, if MariaDB really wants to look Oracle's MySQL face to face, HeidiSQL is not nearly as good as MySQL Workbench. If this wants to be done: 1st) The brand new GUI shall have a Business Source Licence v1.1, which only companies which invest in MariaDB would be able to use. 2nd) In 3 years now, MariaDB would be better positioned and have a fully Open Source GUI which would allow MariaDB to look face to face to Oracle. 3rd) And in 4/5 years now, MariaDB Corporation shall be brave enough to to continue Developing this GUI as a truly open source product, without the BSL, and adopting a purely GPLv2 with incremental improvements, distributing it through the MariaDB Foundation Webpage, as individuals or other companies apart from MariaDB Corporation, will not support the product if it is not totally Open Source. So BSL v1.1 might prove useful to launch the project, but with time, it shall have a GPLv2 for gaining investment from third companies. ¿What do you thing on the topic? Thank you all, Juan Telleria

1 1

[Maria-discuss] MariaDB upgrade 10.1 to 10.2 CentOS7
by Karthick Subramanian 13 Nov '17

13 Nov '17

All, Did anyone face issues after upgrading the mariadb from 10.1 to 10.2 to login to root user as shown below: mysql -u root I got the error - unix_socket not loaded. I noticed that the unix_socket authentication is required after the upgrade for root login from localhost. Can anyone confirm whether this is part of the feature after upgrades. In my previous version: select user, plugin from mysql.user; +-----------+------+-------------+ | Host | User | plugin | +-----------+------+-------------+ | localhost | root | | ------------------------------------ After upgrade: select user, plugin from mysql.user; +-----------+------+-------------+ | Host | User | plugin | +-----------+------+-------------+ | localhost | root | unix_socket | --------------------------------------- Regards, Karthick

3 9

[Maria-discuss] RPM CMake build arguments - Fedora
by Michal Schorm 13 Nov '17

13 Nov '17

Hello everybody, I want to update Fedora SPECfile to the latest best practise. This time, I want to fix CMake arguments. As the time flown, we were adding more and more arguments and for many of them, I'm not sure of their correct values. Are they (the values) ON/OFF? ON/NO? ... I already - incorrectly - filed it as a JIRA issue, but it is nicely formated, so I'd recomment to read it through. https://jira.mariadb.org/browse/MDEV-14344 You can reply to both mail and JIRA, I guess, I'll check them both. -- I'll be glad for any advice on how the values should look like, because reading through the Cmake files is kinda hard and slow for me. -- Michal Schorm Associate Software Engineer Core Services - Databases Team Red Hat

3 4

[Maria-discuss] Contributing my ideas with Mariadb
by Rajat Rawat 11 Nov '17

11 Nov '17

Respected Sir/Mam Hello, Myself Rajat from India. I wanted to contribute/work with Mariadb. As we know that Mariadb dosen’t have cascade truncate function available yet, so I want to contribute to Mariadb towards this topic. My topic is “Cascade Truncate”. Since we can’t truncate the parent table unless the referential integrity constraints are present in the child table also. Hence, we can do a way that we can provide a command that truncate the child table as well. This enhancement is already provided in the Oracle Database in their 12c Version, by using the truncate cascade command. It is better to truncate the child table at the time when we want to truncate our parent table. As we know that we have already on Delete Cascade command available, but that command works only row by row and not for all rows at one time. So, at one time if we want to truncate our whole table then Cascade Truncate is a better Option. That’s all about my proposal. If you like my proposal, then let me know about about it. I am very eager to contribute for you. I also forked and downloaded on my local machine, the file of Maraidb which is available on Github, and currently I am reading/understanding all the patches, but I am confused with only 1 thing that from where should I start. So, if you like proposal then let me know about it and please guide me that from where should I start contributing for it. My skills: C, C++, Java, SQL, PL/SQL Achievement: Oracle Certified Associate Java Programmer SE 7. By contributing for you I will definitely learn many new things and implement my ideas, while working with you. Hope I could be of any help for you. Your sincerely Rajat

2 1

[Maria-discuss] slow_query_log: Binlog Dump
by Reindl Harald 05 Nov '17

05 Nov '17

these entries are *not* helpful when you need to seek for real queryies like a needle SET timestamp=1509853561; # administrator command: Binlog Dump; # Time: 171105 4:47:04 # User@Host: replication[replication] @ [192.168.196.106] # Thread_id: 415842 Schema: QC_hit: No # Query_time: 62.807846 Lock_time: 0.000000 Rows_sent: 1 Rows_examined: 0 # Rows_affected: 0 SET timestamp=1509853624; # administrator command: Binlog Dump; # Time: 171105 4:49:04 # User@Host: replication[replication] @ [192.168.196.106] # Thread_id: 415899 Schema: QC_hit: No # Query_time: 120.228237 Lock_time: 0.000000 Rows_sent: 1 Rows_examined: 0 # Rows_affected: 0 SET timestamp=1509853744; # administrator command: Binlog Dump; # Time: 171105 4:51:01 # User@Host: replication[replication] @ [192.168.196.106] # Thread_id: 416036 Schema: QC_hit: No # Query_time: 117.412204 Lock_time: 0.000000 Rows_sent: 1 Rows_examined: 0 # Rows_affected: 0 SET timestamp=1509853861; # administrator command: Binlog Dump; # Time: 171105 4:52:01 # User@Host: replication[replication] @ [192.168.196.106] # Thread_id: 416184 Schema: QC_hit: No # Query_time: 60.132083 Lock_time: 0.000000 Rows_sent: 1 Rows_examined: 0 # Rows_affected: 0 SET timestamp=1509853921; # administrator command: Binlog Dump; # Time: 171105 4:56:01 # User@Host: replication[replication] @ [192.168.196.106] # Thread_id: 416257 Schema: QC_hit: No # Query_time: 240.021180 Lock_time: 0.000000 Rows_sent: 1 Rows_examined: 0 # Rows_affected: 0 SET timestamp=1509854161; # administrator command: Binlog Dump; # Time: 171105 4:58:08 # User@Host: replication[replication] @ [192.168.196.106] # Thread_id: 416567 Schema: QC_hit: No # Query_time: 127.012674 Lock_time: 0.000000 Rows_sent: 1 Rows_examined: 0 # Rows_affected: 0 SET timestamp=1509854288; # administrator command: Binlog Dump;

1 0

[Maria-discuss] Question on building 10.1.26 binary from sources
by Prasad Rallapalli 03 Nov '17

03 Nov '17

Hi, I am trying to build 10.1.26 from mariadb sources and compare the performance with the release binary for the same query on the same database. cmake using -DBUILD_CONFIG=mysql_release resulted in a "jemalloc not found" error. So, I did a cmake without any parameters and the build was successful. The binary which i built has a performance drop of 20-30% for the queries, compared to that of the released binary, for the same queries. I compared both the mysqld binaries. The release binary seems to be accessing jemalloc functions and the mysqld binary which i built do not have these functions. I am building this on Centos 6.8. Any help or pointers will be appreciated. Regards Prasad

3 3

[Maria-discuss] Building 10.2 from source issues
by Guillaume Lefranc 02 Nov '17

02 Nov '17

Hi, I am trying to build 10.2 from sources while disabling some plugins (building with all plugins works without issues). e.g. $ cmake . -DPLUGIN_MROONGA=NO -DPLUGIN_TOKUDB=NO -DPLUGIN_SPHINX=NO -DPLUGIN_ROCKSDB=NO $ make This fails with the following errors: [ 98%] Building CXX object extra/mariabackup/CMakeFiles/mariabackup.dir/changed_page_bitmap.cc.o In file included from /home/tanj/mariadb-10.2/storage/innobase/include/univ.i:83:0, from /home/tanj/mariadb-10.2/storage/innobase/include/ut0rbt.h:29, from /home/tanj/mariadb-10.2/extra/mariabackup/changed_page_bitmap.h:28, from /home/tanj/mariadb-10.2/extra/mariabackup/changed_page_bitmap.cc:25: /home/tanj/mariadb-10.2/extra/mariabackup/common.h: In function ‘size_t xb_read_full(File, uchar*, size_t)’: /home/tanj/mariadb-10.2/extra/mariabackup/common.h:162:45: error: ‘MY_WME’ was not declared in this scope tbytes = my_read(fd, buf, len - tlen, MYF(MY_WME)); ^ /home/tanj/mariadb-10.2/include/my_global.h:1025:24: note: in definition of macro ‘MYF’ #define MYF(v) (myf) (v) ^ In file included from /home/tanj/mariadb-10.2/extra/mariabackup/changed_page_bitmap.cc:27:0: /home/tanj/mariadb-10.2/extra/mariabackup/common.h:162:12: error: ‘my_read’ was not declared in this scope tbytes = my_read(fd, buf, len - tlen, MYF(MY_WME)); ^~~~~~~ /home/tanj/mariadb-10.2/extra/mariabackup/common.h:162:12: note: suggested alternative: ‘my_sha1’ tbytes = my_read(fd, buf, len - tlen, MYF(MY_WME)); ^~~~~~~ my_sha1 /home/tanj/mariadb-10.2/extra/mariabackup/common.h:163:32: error: ‘MY_FILE_ERROR’ was not declared in this scope if (tbytes == 0 || tbytes == MY_FILE_ERROR) { ^~~~~~~~~~~~~ /home/tanj/mariadb-10.2/extra/mariabackup/common.h:163:32: note: suggested alternative: ‘DB_IO_ERROR’ if (tbytes == 0 || tbytes == MY_FILE_ERROR) { ^~~~~~~~~~~~~ DB_IO_ERROR make[2]: *** [extra/mariabackup/CMakeFiles/mariabackup.dir/build.make:111: extra/mariabackup/CMakeFiles/mariabackup.dir/changed_page_bitmap.cc.o] Error 1 make[2]: *** Waiting for unfinished jobs.... make[1]: *** [CMakeFiles/Makefile2:4270: extra/mariabackup/CMakeFiles/mariabackup.dir/all] Error 2 make[1]: *** Waiting for unfinished jobs.... [ 98%] Built target mysqld [ 98%] Built target explain_filename-t make: *** [Makefile:163: all] Error 2 $ gcc -v Using built-in specs. COLLECT_GCC=gcc COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-pc-linux-gnu/7.2.0/lto-wrapper Target: x86_64-pc-linux-gnu Configured with: /build/gcc-multilib/src/gcc/configure --prefix=/usr --libdir=/usr/lib --libexecdir=/usr/lib --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=https://bugs.archlinux.org/ --enable-languages=c,c++,ada,fortran,go,lto,objc,obj-c++ --enable-shared --enable-threads=posix --enable-libmpx --with-system-zlib --with-isl --enable-__cxa_atexit --disable-libunwind-exceptions --enable-clocale=gnu --disable-libstdcxx-pch --disable-libssp --enable-gnu-unique-object --enable-linker-build-id --enable-lto --enable-plugin --enable-install-libiberty --with-linker-hash-style=gnu --enable-gnu-indirect-function --enable-multilib --disable-werror --enable-checking=release --enable-default-pie --enable-default-ssp Thread model: posix gcc version 7.2.0 (GCC) Please advise. Thanks

2 5

[Maria-discuss] Async replication between two Galera clusters, GTIDs on non slave node in "slave cluster" don't match
by Warner, Paul 02 Nov '17

02 Nov '17

Hi! I have two Galera clusters of three nodes each, let's call them cluster A and cluster B. Cluster B has one member set up as a slave to one of the members in cluster A. The cluster B member that runs the slave protocol shows it's GTID matches that of the nodes in cluster A, but the other two nodes in cluster B (not running slaves themselves, but in the cluster with the node running as a slave) have GTIDs that are always 4 behind the other nodes. So just to lay it out again: A0 5-24-5 - master A1 5-24-5 A2 5-24-5 B0 5-24-5 - slave to A0 B1 5-24-1 B2 5-24-1 The GTIDS stay 4 apart, and replication seems fine. I add rows on any A cluster node and they show up in all three B nodes. The GTID on the node running slave stays in sync with the A cluster. The A0 cluster is set up using xtrabackupv2 for SST. The A cluster is completely up before doing anything on the B cluster. The B cluster is started like: 1. Start B0 as first node in the Galera cluster 2. mysqldump from A0 to B0 with --gtid. <https://gist.github.com/warnerpr/011cabb876600c33e5bee42d7d0b17d2> https://gist.github.com/warnerpr/011cabb876600c33e5bee42d7d0b17d2 3. Master of B0 is set to A0 4. Slave is started on B0 5. B1 and B2 are started and set to join the existing cluster. Cluster A and B have log_slave_updates = ON. Details: B0 Slave - https://gist.github.com/warnerpr/59c287bb201302961742bb89d5035559 B1 - https://gist.github.com/warnerpr/0fb8825e49bce72761e0bdd95b0512d9 A1 Master - https://gist.github.com/warnerpr/91e0df1eb04f425baca3675335639c0f Any pointers are greatly appreciated! Paul

2 1