Gee, and I thought that Bacula is too overwhelming (complicated) with too many directives! :-)
Thanks for the explanation. Kern On 10/03/2016 05:39 PM, Phil Stracchino wrote: > On 10/03/16 10:52, Kern Sibbald wrote: >> On 10/03/2016 02:52 PM, Phil Stracchino wrote: >>> On 10/03/16 05:19, Ralph Bolton wrote: >>>> Yeah, Oracle are making changes alright - not all of them are good :-( >>>> >>>> It's a relatively easy to fix, and actually, it's probably something >>>> that should be changed in the "make_mysql_tables" file supplied with >>>> Bacula (as it's backwards compatible with older MySQL and MariaDBs). The >>>> change is a search and replace for: >>>> >>>> DATETIME DEFAULT 0, >>>> >>>> ...to... >>>> >>>> DATETIME DEFAULT '0000-00-00 00:00:00', >> Hmm. It doesn't surprise me, but I have not run into the problem you >> mention below on any of the MySQL versions that >> are currently in the most popular distros. What MySQL version has that >> change? > > MySQL 5.6 started the stricter enforcement of fully compliant SQL DDL. > However, whether it is enforced or not depends upon a number of MySQL > settings, including SQL_MODE. In 5.5 and older MySQL releases, SQL_MODE > is empty by default, and installations that have been updated in place > will continue to have whatever SQL_MODE was set in the old > configuration, if set. If not set by the configuration files, MySQL > 5.6.6 and later default to NO_ENGINE_SUBSTITUTION. MySQL 5.7.5 added > ONLY_FULL_GROUP_BY and STRICT_TRANS_TABLES to the default SQL_MODE; > MySQL 5.7.7 added NO_AUTO_CREATE_USER; MySQL 5.7.8 added > ERROR_FOR_DIVISION_BY_ZERO, NO_ZERO_DATE, and NO_ZERO_IN_DATE. So a > clean new install of a current MySQL 5.6 release (or an updated one with > no explicitly set SQL_MODE) will use SQL_MODE = NO_ENGINE_SUBSTITUTION, > while a current MySQL 5.7 clean installation (or, again, an updated one > with no explicitly set SQL_MODE) will use NO_ENGINE_SUBSTITUTION, > NO_AUTO_CREATE_USER, ONLY_FULL_GROUP_BY, STRICT_TRANS_TABLES, > NO_ZERO_IN_DATE, NO_ZERO_DATE, ERROR_FOR_DIVISION_BY_ZERO. > > ONLY_FULL_GROUP_BY is not part of the default SQL mode prior to 5.7.5 > because versions prior to 5.7.5 do not detect functional dependency, > which is required to properly implement it. It's not really relevant > here and causes no issues with Bacula. > > NO_ZERO_DATE controls whether '0000-00-00' is considered a valid date. > NO_ZERO_IN_DATE controls whether a DATETIME value is considered valid if > the YEAR is specified but either MONTH or DAY is zero. (For example, > 2016-00-01 or 2016-01-00 are invalid with NO_ZERO_IN_DATE.) Both of > these and ERROR_FOR_DIVISION_BY_ZERO are oddly inconsistent in their > pattern of support. All three were deprecated in 5.7.4, then supported > in 5.7.5 through 5.7.7 but did nothing, supported in 5.7.8 and up but > only do anything in combination with STRICT, and are scheduled to be > deprecated again and removed in some future version when their > functionality is fully rolled into STRICT. > > > I'm currently using: > NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION,STRICT_ALL_TABLES > but NO_AUTO_CREATE USER is also scheduled for future removal (and will > then be on all the time). > > ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
