1. 8
  1.  

  2. 4

    “we do not recommend using it in production”…

    we’ve been running riak 2.0 in production since pre-12. all in all, it has worked out quite well for us and I’m glad we jumped on it early. We’ve had issues, but most of have been of our own creation.

    1. 1

      will you be upgrading to the RC? if you find any future bugs, it would be nice to know if they affect the RC.

      1. 2

        we JUST upgraded to a beta release. Going to RC now seems awfully YOLO. ;)

        1. 1

          i thought you were bleeding edge…

          1. 2

            I don’t think those are mutually exclusive. Sure, running pre-release software means that you have a higher chance of hitting bugs and dealing with feature changes than normal. But whether you are running final versions or pre-release software, you really should be doing solid integration testing to make sure that your software and the third-party package interact as-expected, and that you know, and understand how to work around, any issues that you hit. That integration testing can take time.

            I assume this is exactly the situation SeanTAllen’s in: he’s happy to run pre-release versions, but he wants to do robust testing to make sure that upgrades don’t have unexpected implications. The last round of testing he completed indicated that the beta works great, but he’ll need to repeat that for the RC.

            This is almost exactly the situation Fog Creek was in when I brought in Elasticsearch for Kiln, by the way: we were running it long before it hit 1.0, but we were hardly ever on the newest version, because it took awhile for us to fully vet new versions and make sure we were comfortable with whatever trade-offs were brought with the new release.

            1. 1

              Like having a drunk roman candle fight while wearing eye protection.