October 24, 2014

MySQL and IBM

No, this is not about Sun and IBM :) This is about MySQL. If you download latest 5.1.33 source code you may find there storage/ibmdb2i directory, which obviously is IBM DB2 related. Interesting that there is no mentioning of new engine in Announcement http://dev.mysql.com/doc/refman/5.1/en/news-5-1-33.html.
Quick look into source code says

Also interesting that license of added files is not GPL, but

I think this is outcome of 2-year old press release
“MySQL AB and IBM Announce Open Source Database Support for the IBM System i Platform”, it just took a bit a while to put it into source tree. I wonder what happened with policy not accept significant changes into production release.

About Vadim Tkachenko

Vadim leads Percona's development group, which produces Percona Clould Tools, the Percona Server, Percona XraDB Cluster and Percona XtraBackup. He is an expert in solid-state storage, and has helped many hardware and software providers succeed in the MySQL market.

Comments

  1. alcohol says:

    What a whine post.

    It’s not a change, it’s an addition. Please disconnect from the internet. Thank you.

  2. Rob Wultsch says:

    How cool is that?

    Reminds me of a ./ april fools article a few years back: “Postgres Engine for MySQL Released” http://developers.slashdot.org/article.pl?sid=07/04/01/1448207&from=rss

  3. Vadim says:

    alcohol,

    do you think it is possible PBXT will be included in 5.1 ?
    Feel free to ignore my posts if they do not fit into your internet standards.

  4. Vadim says:

    Rob,

    Not sure how cool is it. Even do not have chance to compile it on System I.
    Maybe joke, but it is in source code :)

  5. This sudden addition of IBMDB2I storage engine may be an indication that IBM may still acquire Sun after all… Incidentally, it seems that an OpenDB2 instance is required in addition to the mentioned IBMDB2I plugin for this plugin to work.

    Speaking of Postgres. Postgres storage engine actually makes more sense than IBMDB2I storage engine.

  6. I filed a bug on this issue: http://bugs.mysql.com/44172

    Very simply, those files either need to be removed from the source tree, or be properly licensed.

  7. erroll says:

    alcohol,
    You’re the whiner – an addition is, obviously, a change. The release has been changed. There’s something in it that wasn’t there before. If you’re going to be arrogant, at least try to be right.

  8. Alexander says:

    i bet this will be soon a bug with license issues in debian, too

  9. The licensing might just be an accident. I remember when I first looked at Falcon in an early release, the license was simply “PROPRIETARY” and when I mentioned it, they fixed it post-haste.

    Putting something new in a production release is an interesting decision. Omitting this from the changelog is even more interesting.

    And still absolutely no word from Sun about this — still more interesting. I’d expect Kaj Arno to blog about it and explain. Perhaps there’s some internal debate going on.

  10. Oh look, I’m denied access to http://bugs.mysql.com/bug.php?id=44172. I noticed that I was denied access to one other bug mentioned in the latest release’s changelog.

    There should be no reason to make bug reports private unless they contain client-confidential data!

  11. William Newton says:

    The mysql bug seems to be publicly accessible again. I’m not sure if I agree with Sun’ legal’s interpretation of the IBM license and the GPL.

    Specifically section B.

    (b) Redistributions in binary form must reproduce this list of conditions, the
    copyright notice in section (d) below, and the disclaimer following this
    list of conditions, in the documentation and/or other materials provided
    with the distribution.

    Isn’t that just a rewording of the BSD advertising clause, which for years made it incompatible with the GPL?

  12. Todd says:

    Not really here to comment on the licensing aspects, but just to note several points:

    0. Shoulda been in the release notes.
    1. This is not a built-in storage engine, it is a plug-in.
    2. The plug-in is in beta/RC state (http://solutions.mysql.com/engines/ibm_db2_storage_engine.html), but will soon reach GA status.
    3. It will be included in both Community and Enterprise releases.
    4. We pushed hard to ensure this was released to the Community first – we want and value community testing.
    5. Packaging with the server is a matter of convenience for System i users.

    I suspect that the argument is being made in the context of, “if they made this change, why not other changes that mean more to me?” I don’t think that’s a meaningful argument in the context of plug-ins, but OK. I’m pretty sure that almost everybody commenting on this doesn’t use or really care about System i, and that – with the exception of the license discussion – nobody really cares that a plug-in for DB2 on that platform is being distributed with Community.

    Too bad that it has to always be a vast conspiracy to undermine the community.

  13. Update…

    License/bug: while the layout is slightly different (initially confusing me), the license appears to be essentially identical to new style BSD (no advertising clause). The old BSD advertising clause was about any other mention of the product (such as advertising) should include the credit – that of course would be unpractical, Berkeley rescinded that long ago and everything is new style BSD these days.
    So all is clear and good now, with thanks to Mark Matthews, Jeffrey Pugh, and Trudy Pelzer.

    @Todd: I don’t care for conspiracies. I agree that adding plugins can be perfectly safe in a release version. I would also say, and you may well agree with this, that adding PBXT as a 5.1 plugin would be the single most useful action that would benefit many.

  14. Vadim says:

    Arjen,

    Thanks for update.

    Re: adding plugin into release:

    As Baron pointed me, it is not perfectly safe. Plugin can have crashes, memory leak, memory corruptions, may initialize server incorrectly, etc

  15. @vadim re plugins. Of course they could cause trouble, but only when they’re loaded which is an explicit/conscious act by the user/DBA.
    The issue is consistency.
    If the DB2 plugin can be added in a release version, so can PBXT – and I would state that PBXT has probably had more real world use already (and potentially so many more users) that it’s likely to have *fewer* bugs that could actually destabilise the server. In that sense it’s quite mature by now.

  16. Todd says:

    Just reiterating the comments I wrote on Baron’s similarly-themed blog:

    1. This is good for community, customers, IBM and Sun – all around.
    2. Plug-ins can crash the server, but only if they are enabled by the user.
    3. The DB2 plugin is the result of a two-year-old partnership between IBM and Sun, not just something we thought would be neat to throw into the server.
    4. We have to be able to line up support, indemnification, training, commercial licenses, etc. that are part of Enterprise offerings and customers need – again, not just something that gets thrown into the server for kicks.
    5. I’m completely, totally, and in all other ways, interested in seeing a similar partnership with other third-party storage engine vendors, such as PBXT. I don’t see it as a double-standard to include the DB2 storage engine when the PBXT relationship hasn’t yet reached the same level.

  17. Vadim says:

    Todd,

    Ok, I think we get to the point where we can close this topic, at least I can’t bring nothing more new here :)

Speak Your Mind

*