Hi Carlos, I read your blog post (actually almost everything I can find on tick tock). My understanding has been tick tock will be the only versioning going forward.
Or are you suggesting at some point there will be a stable train for 3? (or that 3.x will be bumped to 4.0 when stable)? We're on 2.2.5 and haven't seen any major problems with it. -J Sent via iPhone > On Apr 22, 2016, at 03:34, Carlos Rolo <r...@pythian.com> wrote: > > If you need SASI, you need to use 3.4+. 3.x will always be "unstable" (It is > explained why in my blog post). You get those odd versions, but it is not a > solid effort to stabilize the platform, otherwise devs would not jump to 3.6, > and keep working on 3.5. And then you get 3.7, which might fix some issues of > 3.4+, but next month you get 3.8 unstable again... I'm waiting to see where > this is going. I only had bad experiences with 3.x series atm. > > If you want stability (and no new features), you would use 2.1.13. > > 2.2.x is kind of a mixed bag, no really huge improvements over 2.1.x series > and it is still having some issues, so I would stick to 2.1.x series. > > Regards, > > Carlos Juzarte Rolo > Cassandra Consultant / Datastax Certified Architect / Cassandra MVP > > Pythian - Love your data > > rolo@pythian | Twitter: @cjrolo | Linkedin: linkedin.com/in/carlosjuzarterolo > Mobile: +351 91 891 81 00 | Tel: +1 613 565 8696 x1649 > www.pythian.com > >> On Fri, Apr 22, 2016 at 10:16 AM, Jason Williams <jasonjwwilli...@gmail.com> >> wrote: >> My reading of the tick-rock cycle, is that we've moved from a stable train >> that receives mostly bug fixes until the next major stable, to one where >> every odd minor version is a bug fix-only...likely mostly for the previous >> even. The goal being a relatively continuously stable code base in odd minor >> versions. >> >> In that environment where there is no "stable" train, would the right >> approach be to pick the feature set needed and then choose the odd minor >> where that feature set had been stable for 2-3 previous odd minors. >> >> For example, SASI was added in 3.4, so 3.5 is the first bug fix only (odd >> minor) containing it. By the logic above you wouldn't want to use SASI in >> production until 3.9 or later. Or is my logic about how to treat tick-tock >> off base? >> >> -J >> >> >> Sent via iPhone >> >>> On Apr 22, 2016, at 01:46, Satoshi Hikida <sahik...@gmail.com> wrote: >>> >>> Hi, >>> >>> I'm also looking for the most stable version of the Cassandra, too. I read >>> Carlos's blog post. According to his article, I guess 2.1.x is the most >>> stable version, is it right? I prefer to use the most stable version rather >>> than many advanced features. For satisfy my purpose, should I use 2.1.X? or >>> latest 2.2.x is recommended? >>> >>> Currently I use 2.2.5, but is the latest 2.1.13 recommended for production >>> use? >>> >>> Regards, >>> Satoshi >>> >>> >>>> On Mon, Apr 18, 2016 at 11:45 PM, Carlos Rolo <r...@pythian.com> wrote: >>>> Sorry to resurrect this now, but I don't consider anything after 3.0.x >>>> stable. >>>> >>>> I wrote a blog post about this to be clear: >>>> https://www.pythian.com/blog/cassandra-version-production/ >>>> >>>> Use it and pick a version based on your needs. >>>> >>>> Regards, >>>> >>>> Carlos Juzarte Rolo >>>> Cassandra Consultant / Datastax Certified Architect / Cassandra MVP >>>> >>>> Pythian - Love your data >>>> >>>> rolo@pythian | Twitter: @cjrolo | Linkedin: >>>> linkedin.com/in/carlosjuzarterolo >>>> Mobile: +351 91 891 81 00 | Tel: +1 613 565 8696 x1649 >>>> www.pythian.com >>>> >>>>> On Fri, Apr 15, 2016 at 12:44 PM, Jean Tremblay >>>>> <jean.tremb...@zen-innovations.com> wrote: >>>>> Thank you Jack. >>>>> Jean >>>>> >>>>>> On 14 Apr 2016, at 22:00 , Jack Krupansky <jack.krupan...@gmail.com> >>>>>> wrote: >>>>>> >>>>>> Normally, since 3.5 just came out, it would be wise to see if people >>>>>> report any problems over the next few weeks. >>>>>> >>>>>> But... the new tick-tock release process is designed to assure that >>>>>> these odd-numbered releases are only incremental bug fixes from the last >>>>>> even-numbered feature release, which was 3.4. So, 3.5 should be >>>>>> reasonably stable. >>>>>> >>>>>> That said, a bug-fix release of 3.0 is probably going to be more stable >>>>>> than a bug fix release of a more recent feature release (3.4). >>>>>> >>>>>> Usually it comes down to whether you need any of the new features or >>>>>> improvements in 3.x, or whether you might want to keep your chosen >>>>>> release in production for longer than the older 3.0 releases will be in >>>>>> production. >>>>>> >>>>>> Ultimately, this is a personality test: Are you adventuresome or >>>>>> conservative? >>>>>> >>>>>> To be clear, with the new tick-tock release scheme, 3.5 is designed to >>>>>> be a stable release. >>>>>> >>>>>> -- Jack Krupansky >>>>>> >>>>>>> On Thu, Apr 14, 2016 at 3:23 PM, Jean Tremblay >>>>>>> <jean.tremb...@zen-innovations.com> wrote: >>>>>>> Hi, >>>>>>> Could someone give his opinion on this? >>>>>>> What should be considered more stable, Cassandra 3.0.5 or Cassandra 3.5? >>>>>>> >>>>>>> Thank you >>>>>>> Jean >>>>>>> >>>>>>> > On 12 Apr,2016, at 07:00, Jean Tremblay >>>>>>> > <jean.tremb...@zen-innovations.com> wrote: >>>>>>> > >>>>>>> > Hi, >>>>>>> > Which version of Cassandra should considered most stable in the >>>>>>> > version 3? >>>>>>> > I see two main branch: the branch with the version 3.0.* and the >>>>>>> > tick-tock one 3.*.*. >>>>>>> > So basically my question is: which one is most stable, version 3.0.5 >>>>>>> > or version 3.3? >>>>>>> > I know odd versions in tick-took are bug fix. >>>>>>> > Thanks >>>>>>> > Jean >>>> >>>> >>>> -- >>>> > > > -- > > > >
