Mailinglist Archive: opensuse-factory (1264 mails)
| < Previous | Next > |
Re: [opensuse-factory] consolidating the list of mysql flavors in Factory
- From: Dirk Müller <dmueller@xxxxxxx>
- Date: Wed, 14 Sep 2011 10:46:35 +0200
- Message-id: <201109141046.35500.dmueller@suse.de>
On Friday 26 August 2011, Bruno Friedmann wrote:
Hi Bruno,
If I would follow your steps, then that means that at one particular point in
time we have "certified" that it works. What does that help? how does that
help with any fast moving target like openSUSE Factory where things are
constantly being updated, repackaged, etc?
I think focusing on what is maintainable is more important: What can be fixed
easily when it has a problem? Can we get mariadb developers to priotize fixes
for data loss or critical performance issues?
Can we get that for mysql community server? My guess is that for both the
answer is "nay", which means that we can pick one over the other. I don't care
which one, please just make the issue of maintaining a database (and
"supporting" it in the sense of how opensuse is supported) more difficult by
providing multiple flavors of it.
Thanks,
Dirk
--
To unsubscribe, e-mail: opensuse-factory+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse-factory+help@xxxxxxxxxxxx
Dirk to be able to (for example) drop mysql-community version, we need to
test all mysql linked software against mariadb. So first enable a long (I
suspect very long) list of those software. Perhaps there's a hook
somewhere in obs, able to auto-build those deps.
Then test carefully, one by one, to be sure there's no big failure or
regression with mariadb. and notice any changes or optimization needed to
be applied by admins (db conversion etc) in a long readme.
Would be nice to have to support only one. But I've a doubt about the pure
drop'in replacement. It's not like replacing cron by cronie ... There's
user data involved in that case.
Hi Bruno,
If I would follow your steps, then that means that at one particular point in
time we have "certified" that it works. What does that help? how does that
help with any fast moving target like openSUSE Factory where things are
constantly being updated, repackaged, etc?
I think focusing on what is maintainable is more important: What can be fixed
easily when it has a problem? Can we get mariadb developers to priotize fixes
for data loss or critical performance issues?
Can we get that for mysql community server? My guess is that for both the
answer is "nay", which means that we can pick one over the other. I don't care
which one, please just make the issue of maintaining a database (and
"supporting" it in the sense of how opensuse is supported) more difficult by
providing multiple flavors of it.
Thanks,
Dirk
--
To unsubscribe, e-mail: opensuse-factory+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse-factory+help@xxxxxxxxxxxx
| < Previous | Next > |