Hi, "emerge --config percona-server" outputs [...] * Setting root password ... [ ok ] * Loading "zoneinfo", this step may require a few seconds ... ... Warning: Using a password on the command line interface can be insecure. ERROR 1406 (22001) at line 39883: Data too long for column 'Abbreviation' at row 1 [ ok ] * Failed to load zoneinfo! * Stopping the server ... * Done Maybe you only see this error with STRICT SQL modes... Reproducible: Always
http://bugs.mysql.com/bug.php?id=68861
The real culprit is http://bugs.mysql.com/bug.php?id=20545 when STRICT_ALL_TABLES is set before 'emerge --config' is run
I would suggest to replace the "Local time zone must be set--see zic manual page" string with "FACTORY" in mysql-v2.eclass. This way we would end with a "valid" value... it is better than the truncated "Local ti" value, because when you search for "FACTORY timezone" you will find the information that this is a dummy TZ and you have to set a TZ...
Can we make a progress in this bug? For more than 1 year it isn't possible to create a new mysqld out-of-the box on Gentoo due to this problem. At least this workaround (rm /usr/share/zoneinfo/Factory; emerge --config ..; emerge -1 sys-libs/timezone-data) should be documented somewhere... A more convenient way for the eclass seems to be to use /usr/bin/mysql_tzinfo_to_sql to generate the SQL file and sed out the invalid timezones and finally import timezones from the generated file.
(In reply to Thomas D. from comment #4) > Can we make a progress in this bug? For more than 1 year it isn't possible > to create a new mysqld out-of-the box on Gentoo due to this problem. At > least this workaround (rm /usr/share/zoneinfo/Factory; emerge --config ..; > emerge -1 sys-libs/timezone-data) should be documented somewhere... > > A more convenient way for the eclass seems to be to use > /usr/bin/mysql_tzinfo_to_sql to generate the SQL file and sed out the > invalid timezones and finally import timezones from the generated file. I personally don't think there is a need to sed for one user who sets sql_mode=STRICT before system configuration. I'll look into emptying the sql_mode instead for just the init