Comment # 3 on bug 1080891 from Kristyna Streitova

(In reply to Daniel Black from comment #0)

Hello Daniel, thanks for reaching us regarding our patches.

> MariaDB team is trying to take some of the distro patches and is waiting on
> feedback in https://jira.mariadb.org/browse/MDEV-8379.

I've just replied there. Not sure now if you want to discuss the patches here
or in Jira. But I guess that we can discuss the current patches from
server:database here and then you can just update Description in Jira according
to the discussion results. Is it ok for you?


> mariadb-10.0.15-logrotate-su.patch - I like concept, seems possible to
> submit for 10.3

Great!


> mariadb-10.1.1-mysqld_multi-features.patch - I've significantly worked the
> multi-instance systemd service file -
> https://github.com/MariaDB/server/pull/510 - would you object to
> mysqld_multi.sh being dropped in 10.3?

I think that it's ok to drop it in 10.3. Now, we are using our own systemd
solution for starting/stopping both single and multiple instances (see
mysql-systemd-helper [1]). But as mysqld_multi is still shipped we would like
to keep this patch till it's dropped by upstream as users can rely on this
additional features. When updating to 10.3 we should also review our helper
script and maybe we could move closer to upstream.


> mariadb-10.1.12-deharcode-libdir.patch - will look at later. Probably could
> do a SUSE layout in some way.

Ok, thanks!


> mariadb-10.2.10-disable_openssl_compatibility_check.patch - fixed
> https://github.com/MariaDB/server/pull/558

I can confirm that it's fixed.


> mariadb-10.2.3-group.patch - was per MDEV-8379

I responded in MDEV-8379. Basically, the patch looks like a half solution and
we will probably drop it. However, the idea of adding support for group option
is nice I guess.


> mariadb-10.2.4-logrotate.patch - looks good - @localstatedir@ need
> examination as to how to be distro-agnostic.

In MDEV-8379 Sergei mentioned that the path can be configurable and then we
would need to build it with something like DLOG_PATH=/var/log/mysql. I think
that this compilation option is definitely better than a patch for us. So we
are ok with it.


> mariadb-10.2.8-systemd-cmake.patch - not sure why handling so many options
> is needed - looks like -DWITH_SYSTEMD=no would follow the true paths

I did some testing regarding this patch. It seems that when we leave
WITH_SYSTEMD option out, we can remove this patch and the systemd detection
works just fine for us.


> mariadb-10.2.9-galera_cnf.patch - possible, if I can get the provider path
> correct in a distro agnostic way.

It would be great to upstream this patch as I think that it's useful for users
that are not so familiar with the galera configuration.


> mariadb-5.2.3-cnf.patch - all these files have been removed in 10.3

Ok, we will remove it when we update to 10.3.


> mariadb-5.5.28-install_db-quiet - I've seen response in MDEV-8379, I see
> this is called from systemd service file. I'll see what can be done.

Thanks! 


[1]
https://build.opensuse.org/package/view_file/openSUSE:Factory/mariadb/mysql-systemd-helper?expand=1