Re: [openstack-dev] [openstack-ansible] Change of default database collation

2016-06-24 Thread Major Hayden 
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256

On 06/24/2016 12:07 AM, Jimmy McCrory wrote:
> The question then comes to how to best handle upgrades from Mitaka to Newton.
> Any input for the current proposal[3] from anyone that may have experience 
> with any project's database migration scripts, or MySQL-based databases in 
> general, would be appreciated.

Your proposal looks fine, and from what I read, it seems like the difference 
between utf8_unicode and utf8_general is mainly around sorting Unicode 
characters.  Then again, I'm no collation/unicode expert, so I'll gladly defer 
to anyone with more experience there. ;)

Do we know if a change from utf8_unicode_ci to utf8_general_ci will change data 
in the database?  If it doesn't, I'd imagine that a wholesale conversion to 
utf8_general_ci (as you've proposed in 333733) would be just fine.  There are 
some small performance concerns documented[1], but they don't look huge.

[1] 
http://stackoverflow.com/questions/766809/whats-the-difference-between-utf8-general-ci-and-utf8-unicode-ci

- --
Major Hayden
-BEGIN PGP SIGNATURE-
Version: GnuPG v2
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=63dN
-END PGP SIGNATURE-

__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

[openstack-dev] [openstack-ansible] Change of default database collation

2016-06-23 Thread Jimmy McCrory 
Hi all,

OpenStack-Ansible has been configuring the default database collation as
'utf8_unicode_ci'.
We've recently run into an issue with new deployments during cinder
database migrations where a foreign key column had different a collation
between the parent and child table[1].

That's since been fixed, and we're now looking at changing the default
collation to match the default of MySQL/MariaDB's utf8 character set to
avoid the possibility of this same discrepancy with new deployments[2].

The question then comes to how to best handle upgrades from Mitaka to
Newton.
Any input for the current proposal[3] from anyone that may have experience
with any project's database migration scripts, or MySQL-based databases in
general, would be appreciated.

Also,
Have any of the other deployment projects been effected by this?
Are there any in-progress efforts to help further enforce a standard
character set and collation through DB migration scripts?

Thanks

[1] https://bugs.launchpad.net/cinder/+bug/1594195
[2] https://review.openstack.org/#/c/331786/
[3] https://review.openstack.org/#/c/333733/
__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev