Re: [Maria-discuss] transactions and UNLOCK TABLES
I would never use LOCK TABLES on InnoDB, look again my first test case and you'll see that it's different from what documentation says :) Federico Mar 23/9/14, Roberto Spadim robe...@spadim.com.br ha scritto: Oggetto: Re: [Maria-discuss] transactions and UNLOCK TABLES A: Federico Razzoli federico_...@yahoo.it Cc: Maria Discuss maria-discuss@lists.launchpad.net Data: Martedì 23 settembre 2014, 01:46 well from kb what i understand is:to use unlock with innodb (the engine you use with both tables), you should SET innodb_table_locks=1; SET autocommit=0; after that any LOCK/UNLOCK with innodb tables inside a transaction will commit the transaction try again using innodb_table_locks, and autocommit variables 2014-09-22 19:51 GMT-03:00 Federico Razzoli federico_...@yahoo.it: No, becuase the first test case I've posted shows that the reality is different... :) I'm just asking to explain the exact interaction between trx and UNLOCK, because I need to know if statements I use are safe. Federico Mar 23/9/14, Roberto Spadim robe...@spadim.com.br ha scritto: Oggetto: Re: [Maria-discuss] transactions and UNLOCK TABLES A: Federico Razzoli federico_...@yahoo.it Cc: Maria Discuss maria-discuss@lists.launchpad.net, maria-docs maria-d...@lists.launchpad.net Data: Martedì 23 settembre 2014, 00:14 better explained at mariadb kbhttps://mariadb.com/kb/en/mariadb/documentation/sql-commands/transactions/lock-tables-and-unlock-tables/#limitations ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp -- Roberto Spadim SPAEmpresarialEng. Automação e Controle ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
[Maria-discuss] R: transactions and UNLOCK TABLES
Let's make the snippet simpler. Only one table, non-transactional, please look at the difference from what docs say and the real behaviour. If one could explain what UNLOCK TABLES exactly does, it would be great. MariaDB [test] SELECT @@in_transaction, @@autocommit; +--+--+ | @@in_transaction | @@autocommit | +--+--+ |0 |1 | +--+--+ 1 row in set (0.00 sec) MariaDB [test] CREATE OR REPLACE TABLE t (c INT) ENGINE = MEMORY; Query OK, 0 rows affected (0.12 sec) MariaDB [test] LOCK TABLE t WRITE; Query OK, 0 rows affected (0.00 sec) MariaDB [test] START TRANSACTION; Query OK, 0 rows affected (0.00 sec) MariaDB [test] INSERT INTO t VALUES (1); Query OK, 1 row affected (0.01 sec) MariaDB [test] UNLOCK TABLES; Query OK, 0 rows affected (0.00 sec) MariaDB [test] SELECT @@in_transaction; +--+ | @@in_transaction | +--+ |1 | +--+ 1 row in set (0.00 sec) Regards Federico ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] R: transactions and UNLOCK TABLES
Hi Federico, On 23.09.2014 12:50, Federico Razzoli wrote: Let's make the snippet simpler. Only one table, non-transactional, please look at the difference from what docs say and the real behaviour. If one could explain what UNLOCK TABLES exactly does, it would be great. MariaDB [test] SELECT @@in_transaction, @@autocommit; +--+--+ | @@in_transaction | @@autocommit | +--+--+ |0 |1 | +--+--+ 1 row in set (0.00 sec) MariaDB [test] CREATE OR REPLACE TABLE t (c INT) ENGINE = MEMORY; Query OK, 0 rows affected (0.12 sec) MariaDB [test] LOCK TABLE t WRITE; Query OK, 0 rows affected (0.00 sec) MariaDB [test] START TRANSACTION; Query OK, 0 rows affected (0.00 sec) Beginning a transaction causes table locks acquired with LOCK TABLES to be released, as though you had executed UNLOCK TABLES. http://dev.mysql.com/doc/refman/5.5/en/commit.html MariaDB [test] INSERT INTO t VALUES (1); Query OK, 1 row affected (0.01 sec) MariaDB [test] UNLOCK TABLES; Query OK, 0 rows affected (0.00 sec) So, this UNLOCK is not actually doing anything. Regards, Elena MariaDB [test] SELECT @@in_transaction; +--+ | @@in_transaction | +--+ |1 | +--+ 1 row in set (0.00 sec) Regards Federico ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] R: transactions and UNLOCK TABLES
Thank you for your clarification, Elena. Based on this information, I find the MySQL manual confusing, so I have reported a bug: http://bugs.mysql.com/bug.php?id=74033 Unless it turns out that I'm still missing something, I will also check the kb. Regards Federico Mar 23/9/14, Elena Stepanova ele...@montyprogram.com ha scritto: Oggetto: Re: [Maria-discuss] R: transactions and UNLOCK TABLES A: Federico Razzoli federico_...@yahoo.it, maria-discuss@lists.launchpad.net Data: Martedì 23 settembre 2014, 11:51 Hi Federico, On 23.09.2014 12:50, Federico Razzoli wrote: Let's make the snippet simpler. Only one table, non-transactional, please look at the difference from what docs say and the real behaviour. If one could explain what UNLOCK TABLES exactly does, it would be great. MariaDB [test] SELECT @@in_transaction, @@autocommit; +--+--+ | @@in_transaction | @@autocommit | +--+--+ | 0 | 1 | +--+--+ 1 row in set (0.00 sec) MariaDB [test] CREATE OR REPLACE TABLE t (c INT) ENGINE = MEMORY; Query OK, 0 rows affected (0.12 sec) MariaDB [test] LOCK TABLE t WRITE; Query OK, 0 rows affected (0.00 sec) MariaDB [test] START TRANSACTION; Query OK, 0 rows affected (0.00 sec) Beginning a transaction causes table locks acquired with LOCK TABLES to be released, as though you had executed UNLOCK TABLES. http://dev.mysql.com/doc/refman/5.5/en/commit.html MariaDB [test] INSERT INTO t VALUES (1); Query OK, 1 row affected (0.01 sec) MariaDB [test] UNLOCK TABLES; Query OK, 0 rows affected (0.00 sec) So, this UNLOCK is not actually doing anything. Regards, Elena MariaDB [test] SELECT @@in_transaction; +--+ | @@in_transaction | +--+ | 1 | +--+ 1 row in set (0.00 sec) Regards Federico ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] R: transactions and UNLOCK TABLES
+1 =) mysql manual sometimes don't help, that's why i really prefer mariadb kb i think that many years ago mysql implement this nontransactional like locks to innodb just to allow guys porting from myisam to innodb using old apps without rewriting the lock/unlock to begin/commit if you are using innodb as transactional you will never use lock/unlock, in my opnion maybe you probably only use locks at filesystem or use get_lock() mysql function or any other IPC function, but be carefull about scalability i don't know what's the 'best' lock server today, all shared locks that i use today i use over a filesystem and network (nfs/smb/others network file system), if get_lock work with galera maybe i could get a try 2014-09-23 6:51 GMT-03:00 Elena Stepanova ele...@montyprogram.com: Hi Federico, On 23.09.2014 12:50, Federico Razzoli wrote: Let's make the snippet simpler. Only one table, non-transactional, please look at the difference from what docs say and the real behaviour. If one could explain what UNLOCK TABLES exactly does, it would be great. MariaDB [test] SELECT @@in_transaction, @@autocommit; +--+--+ | @@in_transaction | @@autocommit | +--+--+ |0 |1 | +--+--+ 1 row in set (0.00 sec) MariaDB [test] CREATE OR REPLACE TABLE t (c INT) ENGINE = MEMORY; Query OK, 0 rows affected (0.12 sec) MariaDB [test] LOCK TABLE t WRITE; Query OK, 0 rows affected (0.00 sec) MariaDB [test] START TRANSACTION; Query OK, 0 rows affected (0.00 sec) Beginning a transaction causes table locks acquired with LOCK TABLES to be released, as though you had executed UNLOCK TABLES. http://dev.mysql.com/doc/refman/5.5/en/commit.html MariaDB [test] INSERT INTO t VALUES (1); Query OK, 1 row affected (0.01 sec) MariaDB [test] UNLOCK TABLES; Query OK, 0 rows affected (0.00 sec) So, this UNLOCK is not actually doing anything. Regards, Elena MariaDB [test] SELECT @@in_transaction; +--+ | @@in_transaction | +--+ |1 | +--+ 1 row in set (0.00 sec) Regards Federico ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] R: transactions and UNLOCK TABLES
nice, i will include the information about lock innodb tables to your bug request just to help a better explain at mysql docs 2014-09-23 8:50 GMT-03:00 Federico Razzoli federico_...@yahoo.it: Thank you for your clarification, Elena. Based on this information, I find the MySQL manual confusing, so I have reported a bug: http://bugs.mysql.com/bug.php?id=74033 Unless it turns out that I'm still missing something, I will also check the kb. Regards Federico Mar 23/9/14, Elena Stepanova ele...@montyprogram.com ha scritto: Oggetto: Re: [Maria-discuss] R: transactions and UNLOCK TABLES A: Federico Razzoli federico_...@yahoo.it, maria-discuss@lists.launchpad.net Data: Martedì 23 settembre 2014, 11:51 Hi Federico, On 23.09.2014 12:50, Federico Razzoli wrote: Let's make the snippet simpler. Only one table, non-transactional, please look at the difference from what docs say and the real behaviour. If one could explain what UNLOCK TABLES exactly does, it would be great. MariaDB [test] SELECT @@in_transaction, @@autocommit; +--+--+ | @@in_transaction | @@autocommit | +--+--+ | 0 |1 | +--+--+ 1 row in set (0.00 sec) MariaDB [test] CREATE OR REPLACE TABLE t (c INT) ENGINE = MEMORY; Query OK, 0 rows affected (0.12 sec) MariaDB [test] LOCK TABLE t WRITE; Query OK, 0 rows affected (0.00 sec) MariaDB [test] START TRANSACTION; Query OK, 0 rows affected (0.00 sec) Beginning a transaction causes table locks acquired with LOCK TABLES to be released, as though you had executed UNLOCK TABLES. http://dev.mysql.com/doc/refman/5.5/en/commit.html MariaDB [test] INSERT INTO t VALUES (1); Query OK, 1 row affected (0.01 sec) MariaDB [test] UNLOCK TABLES; Query OK, 0 rows affected (0.00 sec) So, this UNLOCK is not actually doing anything. Regards, Elena MariaDB [test] SELECT @@in_transaction; +--+ | @@in_transaction | +--+ |1 | +--+ 1 row in set (0.00 sec) Regards Federico ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] OQGRAPH and query cache
Hi! On 23.09.14 17:25, Roberto Spadim wrote: Hi guys, i was reading docs, and found this: * The query cache needs to be disabled in order to use OQGRAPH https://mariadb.com/kb/en/oqgraph/ and, in older versions, Galera https://mariadb.com/kb/en/galera/. Why query cache needs to be disabled with OQGRAPH? My guess is that because it can't cooperate with QC and behaviour to not match our 2 defaukt patterns (like myisam or transactional). If it fit in one of the patterns it could be switched on. [skip] ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] OQGRAPH and query cache
i'm thinking about how it work with backend data store for example... select somefiled from oq_graph where something.. oq_graph will send some queries to data table (i don't know if it use sql or internal handler) data table will return oq_graph will process and send back data to client, here it can/can't query cache if data table change, oq_graph don't know and old query cache will become inconsistent, the point is, why not it need to be disabled? should be nice a explain about it, i don't know too how it works :/ any other idea? 2014-09-23 12:25 GMT-03:00 Oleksandr Byelkin sa...@montyprogram.com: Hi! On 23.09.14 17:25, Roberto Spadim wrote: Hi guys, i was reading docs, and found this: * The query cache needs to be disabled in order to use OQGRAPH https://mariadb.com/kb/en/oqgraph/ and, in older versions, Galera https://mariadb.com/kb/en/galera/. Why query cache needs to be disabled with OQGRAPH? My guess is that because it can't cooperate with QC and behaviour to not match our 2 defaukt patterns (like myisam or transactional). If it fit in one of the patterns it could be switched on. [skip] -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] OQGRAPH and query cache
other doubt https://github.com/MariaDB/server/blob/10.1/storage/oqgraph/ha_oqgraph.cc#L462 when using aria as backstorage, aria engine don't support insert delayed, this will make some problems not? about query cache i didn't found a register_query_cache_table function like aria/innodb/myisam engines, maybe we should change documentation to 'query cache isn't supported with oqgraph' instead of 'needs to be disabled' 2014-09-23 14:13 GMT-03:00 Roberto Spadim robe...@spadim.com.br: i'm thinking about how it work with backend data store for example... select somefiled from oq_graph where something.. oq_graph will send some queries to data table (i don't know if it use sql or internal handler) data table will return oq_graph will process and send back data to client, here it can/can't query cache if data table change, oq_graph don't know and old query cache will become inconsistent, the point is, why not it need to be disabled? should be nice a explain about it, i don't know too how it works :/ any other idea? 2014-09-23 12:25 GMT-03:00 Oleksandr Byelkin sa...@montyprogram.com: Hi! On 23.09.14 17:25, Roberto Spadim wrote: Hi guys, i was reading docs, and found this: * The query cache needs to be disabled in order to use OQGRAPH https://mariadb.com/kb/en/oqgraph/ and, in older versions, Galera https://mariadb.com/kb/en/galera/. Why query cache needs to be disabled with OQGRAPH? My guess is that because it can't cooperate with QC and behaviour to not match our 2 defaukt patterns (like myisam or transactional). If it fit in one of the patterns it could be switched on. [skip] -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] Spider and query cache
hi kentoku! i will try with mysql server but i got one question, why mariadb don't support VP? is something different that block using VP ? 2014-09-23 14:44 GMT-03:00 kentoku kentokush...@gmail.com: Hi Roberto, Thank you for testing VP. I'll think about this feature. Thanks, Kentoku 2014-09-21 18:10 GMT+09:00 Roberto Spadim robe...@spadim.com.br: Hi! Well i'm checking vp isn't part of mariadb today :/ maybe we should port it first and after jmplement new features, right? Em domingo, 21 de setembro de 2014, Oleksandr Byelkin sa...@montyprogram.com escreveu: Hi! On 21.09.14 05:06, Roberto Spadim wrote: [skip] sounds crazy? :) ok i know that query cache have some problems with mutex, but it's another history we could have a better hit rate with this feature (at least when correctly used) check that query cache become inconsistent if we change data1 or data2 outside vp table It is doable if put a lot of hooks of asking engine about tables. There is already hooks now used in table invalidation. But it works slowly (relatively) because having no open table difficult to give engine full information and so each such hook will cause semi-open process and so if we have miss it looks like double opening overhead. So, in price of total overhead we can make it, the question is if it help so much that gain will be more then overhead in average... As usual I have doubts about it (need a proof). [skip] -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] Spider and query cache
https://launchpad.net/vpformysql/trunk/1.1-for-5.5.34-10.0.9/+download/vp-src-1.1-for-5.5.34-10.0.9.tgz now i saw 10.0.9 at 'last' name, i will try it :) news i post here 2014-09-23 14:54 GMT-03:00 Roberto Spadim robe...@spadim.com.br: hi kentoku! i will try with mysql server but i got one question, why mariadb don't support VP? is something different that block using VP ? 2014-09-23 14:44 GMT-03:00 kentoku kentokush...@gmail.com: Hi Roberto, Thank you for testing VP. I'll think about this feature. Thanks, Kentoku 2014-09-21 18:10 GMT+09:00 Roberto Spadim robe...@spadim.com.br: Hi! Well i'm checking vp isn't part of mariadb today :/ maybe we should port it first and after jmplement new features, right? Em domingo, 21 de setembro de 2014, Oleksandr Byelkin sa...@montyprogram.com escreveu: Hi! On 21.09.14 05:06, Roberto Spadim wrote: [skip] sounds crazy? :) ok i know that query cache have some problems with mutex, but it's another history we could have a better hit rate with this feature (at least when correctly used) check that query cache become inconsistent if we change data1 or data2 outside vp table It is doable if put a lot of hooks of asking engine about tables. There is already hooks now used in table invalidation. But it works slowly (relatively) because having no open table difficult to give engine full information and so each such hook will cause semi-open process and so if we have miss it looks like double opening overhead. So, in price of total overhead we can make it, the question is if it help so much that gain will be more then overhead in average... As usual I have doubts about it (need a proof). [skip] -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] Spider and query cache
hi kentoku! i will try with mysql server but i got one question, why mariadb don't support VP? is something different that block using VP ? It is need some changes for MariaDB. now i saw 10.0.9 at 'last' name, i will try it :) Please use this for now. source code http://spiderformysql.com/downloads/spider-3.2/mariadb-10.0.13-spider-3.2-vp-1.1-mroonga-4.05a.tgz binary for linux x86_64 http://spiderformysql.com/downloads/spider-3.2/mariadb-10.0.13-spider-3.2-vp-1.1-mroonga-4.05-linux-x86_64-glibc25a.tgz Thanks, Kentoku 2014-09-24 3:06 GMT+09:00 Roberto Spadim robe...@spadim.com.br: https://launchpad.net/vpformysql/trunk/1.1-for-5.5.34-10.0.9/+download/vp-src-1.1-for-5.5.34-10.0.9.tgz now i saw 10.0.9 at 'last' name, i will try it :) news i post here 2014-09-23 14:54 GMT-03:00 Roberto Spadim robe...@spadim.com.br: hi kentoku! i will try with mysql server but i got one question, why mariadb don't support VP? is something different that block using VP ? 2014-09-23 14:44 GMT-03:00 kentoku kentokush...@gmail.com: Hi Roberto, Thank you for testing VP. I'll think about this feature. Thanks, Kentoku 2014-09-21 18:10 GMT+09:00 Roberto Spadim robe...@spadim.com.br: Hi! Well i'm checking vp isn't part of mariadb today :/ maybe we should port it first and after jmplement new features, right? Em domingo, 21 de setembro de 2014, Oleksandr Byelkin sa...@montyprogram.com escreveu: Hi! On 21.09.14 05:06, Roberto Spadim wrote: [skip] sounds crazy? :) ok i know that query cache have some problems with mutex, but it's another history we could have a better hit rate with this feature (at least when correctly used) check that query cache become inconsistent if we change data1 or data2 outside vp table It is doable if put a lot of hooks of asking engine about tables. There is already hooks now used in table invalidation. But it works slowly (relatively) because having no open table difficult to give engine full information and so each such hook will cause semi-open process and so if we have miss it looks like double opening overhead. So, in price of total overhead we can make it, the question is if it help so much that gain will be more then overhead in average... As usual I have doubts about it (need a proof). [skip] -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] Spider and query cache
hum, i'm reading source... some changes to handler and something at sql_class related to xid_? plugins, fields soon i will compile and test ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
Re: [Maria-discuss] OQGRAPH and query cache
well found this bug at mariadb jira https://mariadb.atlassian.net/browse/MDEV-5744 that's why query cache should be disabled, but check that it's a problem of oqgraph engine, that's a BUG! 2014-09-23 15:41 GMT-03:00 Peter Laursen peter_laur...@webyog.com: INSERT DELAYED is on its way _out _in_the_dark anyway. I blogged this http://blog.webyog.com/2012/08/01/insert-delayed-is-gone/ 2 years ago. I am not able to really to recall full details from my brain-dead mind, but I think all references to INSERT DELAYED are now uninteresting. -- Peter -- Webyog On Tue, Sep 23, 2014 at 7:51 PM, Roberto Spadim robe...@spadim.com.br wrote: other doubt https://github.com/MariaDB/server/blob/10.1/storage/oqgraph/ha_oqgraph.cc#L462 when using aria as backstorage, aria engine don't support insert delayed, this will make some problems not? about query cache i didn't found a register_query_cache_table function like aria/innodb/myisam engines, maybe we should change documentation to 'query cache isn't supported with oqgraph' instead of 'needs to be disabled' 2014-09-23 14:13 GMT-03:00 Roberto Spadim robe...@spadim.com.br: i'm thinking about how it work with backend data store for example... select somefiled from oq_graph where something.. oq_graph will send some queries to data table (i don't know if it use sql or internal handler) data table will return oq_graph will process and send back data to client, here it can/can't query cache if data table change, oq_graph don't know and old query cache will become inconsistent, the point is, why not it need to be disabled? should be nice a explain about it, i don't know too how it works :/ any other idea? 2014-09-23 12:25 GMT-03:00 Oleksandr Byelkin sa...@montyprogram.com: Hi! On 23.09.14 17:25, Roberto Spadim wrote: Hi guys, i was reading docs, and found this: * The query cache needs to be disabled in order to use OQGRAPH https://mariadb.com/kb/en/oqgraph/ and, in older versions, Galera https://mariadb.com/kb/en/galera/. Why query cache needs to be disabled with OQGRAPH? My guess is that because it can't cooperate with QC and behaviour to not match our 2 defaukt patterns (like myisam or transactional). If it fit in one of the patterns it could be switched on. [skip] -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp -- Roberto Spadim SPAEmpresarial Eng. Automação e Controle ___ Mailing list: https://launchpad.net/~maria-discuss Post to : maria-discuss@lists.launchpad.net Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp