On Wed, May 30, 2018 at 11:30 PM, Kevin Kofler <kevin.kofler(a)chello.at>
wrote:
Michal Schorm wrote:
> So far, on F>=28 the TokuDB was build without jemalloc.
> It may be risky to rely on, yet no one complained.
IMHO, if it works (well enough that nobody is complaining), just keep
shipping it that way. It's better than to break dependencies (if you don't
Obsolete the subpackage) or silently drop functionality (if you do) in an
update (which will probably even be a security update) to a stable release.
The biggest problem for me is, that as a maintainer, I have no idea, how
wide, skilled or satisfied the user base is.
If something is obviously broken, I get bugzilla filed or Bodhi update
commented.
But other than that, I don't know if - for example in TokuDB - are problems
present, but there is some workarounds, which skilled DB admins use but it
doesn work for them by default.
So I just hope "silence means yes - it works".
--
Michal Schorm
Associate Software Engineer
Core Services - Databases Team
Red Hat