Announcement

Collapse
No announcement yet.

A Few Worrisome Regressions Appear In Ubuntu 15.04 vs. 15.10 Performance

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #51
    Originally posted by gQuigs View Post
    Tried it in two VMs (on OpenStack) and didn't find a significant difference. http://openbenchmarking.org/result/1...BE-1510118BE07
    VM != Hardware when it comes to benchmarking. Especially if you're not the only person on that host.

    Comment


    • #52
      Originally posted by Mike Frett View Post
      Don't they say that Slackware is the only true Linux? =p
      Arch sux.

      Ubuntu sux.

      The one true Linux is Xenix.
      https://en.wikipedia.org/wiki/Xenix

      Comment


      • #53
        Originally posted by SystemCrasher View Post
        And btw, real software engineers would not tolerate system breakages in production.
        too bad there are so many fake engineers and impostors around....

        Comment


        • #54
          Originally posted by nils_ View Post

          too bad there are so many fake engineers and impostors around....
          Sure. And engineer is not one who yells most. IMHO, one can call self an engineer when others start calling you an engineer and treat you like if you were an engineer in some field of expertise, asking for advice and being sure you'll do your best to find best possible solution for their problem, etc. That's when one can start considering self an engineer and even likely to be recognized by others as such. And just yelling some unbacked bold claims on forum does not makes one an engineer.

          Comment


          • #55
            Originally posted by SystemCrasher View Post
            Sure. And engineer is not one who yells most. IMHO, one can call self an engineer when others start calling you an engineer and treat you like if you were an engineer in some field of expertise, asking for advice and being sure you'll do your best to find best possible solution for their problem, etc. That's when one can start considering self an engineer and even likely to be recognized by others as such. And just yelling some unbacked bold claims on forum does not makes one an engineer.
            No true Scotsman.

            Comment


            • #56
              [QUOTE=SystemCrasher;n828600]
              /QUOTE]
              Gentoo stable moves actually quite slow. It is often behind Ubuntu with releases of libraries etc. Regarding testing, well you can anyway never test all combos, this is about man power etc., but Gentroo developers do their best I think to test stable branch, and especially the core of the system, server packages and such.
              I am primarily Ubuntu user myself during last years, but I was using Gentoo for ~5 years, on different systems, and although some things are definitely more simple with Ubuntu, or just work, I don't think claiming Ubuntu is more stable than Gentoo (stable) makes any sense. Even unstable branch is very stable, probably more stable than Arch.
              Great thing with Gentoo is than one can mix branches (stay on stable, while only using few packages from unstable), and also unmask (masked) packages if one really wants it. If system goes kaboom one day, portage can fix it.
              I don't say Gentoo doesn't have its issues, but I don't think 'stability' is one of them.

              Comment


              • #57
                Originally posted by reCAPTCHA View Post
                Gentoo stable moves actually quite slow. It is often behind Ubuntu with releases of libraries etc.
                And I never seen anyone using it, in first place. When someone tells he/she uses gentoo, you can count on very particular set of properties and "stable" usually put out of equation.

                Regarding testing, well you can anyway never test all combos, this is about man power etc., but Gentroo developers do their best I think to test stable branch, and especially the core of the system, server packages and such.
                I bet Ubuntu beats it to the dust, because of far more users. Same goes for Debian, etc - it just got enough people and plenty time to chew on between releases. RedHat can surely afford to hire enough of full time QAs, devs, etc. So I understand who would test it, when and why and what to expect.

                Then, its about community. Testing your super-unique mega-installation often does not benefits anyone except you. Face that. This combo could be unique on the whole planet. Nobody gives a fuck how it performs. It just irrelevant for anyone else. They are not likely to have same installation anyway. And if they change half of system... they are doomed to have different experience. And different bugs. Since gentoo users are mostly camikaze-inclined, I can imagine only few devs "tested" it. With fairly predictable coverage.

                I don't think claiming Ubuntu is more stable than Gentoo (stable) makes any sense.
                TBH, as far as I understand, whole point of those using Gentoo is ability to get some custom configuration. This inherently assumes they are getting unique and ... almost completely untested thing. If you changed build flags and options, feel free to put all previous testing results into garbage bin and restart testing from scratch. It do not have to be taken as granted that previous test results are anyhow valid in this case.

                Even unstable branch is very stable, probably more stable than Arch.
                More stable than Arch... it sounds interesting, sure. Arch is so "stable" it can deploy you systemd right into middle of running system. And screw it up, if you do not RTFM about each and every update and wouldn't handle rough edges yourself. So, beating Arch isn't a big achievement, to say the least.

                Great thing with Gentoo is than one can mix branches (stay on stable, while only using few packages from unstable),
                One can do something like this on Debian, too. As well as one can add PPAs to ubuntu, etc. That's what users of Oibaf or paulo diaz PPAs are doing, etc. So it's not like if it some unique feature of Gentoo. But sure, different approaches and tradeoffs have different sets of resulting advantages and drawbacks.

                and also unmask (masked) packages if one really wants it.
                You tell it like if other package management tools are nuts and can't do it. This is not true, to say the least. But sure, it takes some learning to do it, and it may or may not fit particular use case better.

                I don't say Gentoo doesn't have its issues, but I don't think 'stability' is one of them.
                On other hand, from my experience, false alarms in bug trackers are hallmark of Gentoo users. So I wouldn't agree your views on this topic, sorry.

                Comment


                • #58
                  Originally posted by SystemCrasher View Post
                  And I never seen anyone using it, in first place. When someone tells he/she uses gentoo, you can count on very particular set of properties and "stable" usually put out of equation.

                  I bet Ubuntu beats it to the dust, because of far more users. Same goes for Debian, etc - it just got enough people and plenty time to chew on between releases. RedHat can surely afford to hire enough of full time QAs, devs, etc. So I understand who would test it, when and why and what to expect.

                  Then, its about community. Testing your super-unique mega-installation often does not benefits anyone except you. Face that. This combo could be unique on the whole planet. Nobody gives a fuck how it performs. It just irrelevant for anyone else. They are not likely to have same installation anyway. And if they change half of system... they are doomed to have different experience. And different bugs. Since gentoo users are mostly camikaze-inclined, I can imagine only few devs "tested" it. With fairly predictable coverage.


                  TBH, as far as I understand, whole point of those using Gentoo is ability to get some custom configuration. This inherently assumes they are getting unique and ... almost completely untested thing. If you changed build flags and options, feel free to put all previous testing results into garbage bin and restart testing from scratch. It do not have to be taken as granted that previous test results are anyhow valid in this case.


                  More stable than Arch... it sounds interesting, sure. Arch is so "stable" it can deploy you systemd right into middle of running system. And screw it up, if you do not RTFM about each and every update and wouldn't handle rough edges yourself. So, beating Arch isn't a big achievement, to say the least.


                  One can do something like this on Debian, too. As well as one can add PPAs to ubuntu, etc. That's what users of Oibaf or paulo diaz PPAs are doing, etc. So it's not like if it some unique feature of Gentoo. But sure, different approaches and tradeoffs have different sets of resulting advantages and drawbacks.


                  You tell it like if other package management tools are nuts and can't do it. This is not true, to say the least. But sure, it takes some learning to do it, and it may or may not fit particular use case better.


                  On other hand, from my experience, false alarms in bug trackers are hallmark of Gentoo users. So I wouldn't agree your views on this topic, sorry.
                  Are you aware that NASDAQ uses Gentoo, so as Google (ie for Chrome OS, or that router from Phoronix news.), NASA for exploration rovers, lot of scientific projects use it, and it is really great system for appliances.
                  Way you compared Gentoo with Debian/Ubuntu when we discuss branches and package manager (Mixing, unmasking packages etc.) indicates you actually have little experience on the topic, otherwise you wouldn't type something like "And I never seen anyone using it, in first place. When someone tells he/she uses gentoo, you can count on very particular set of properties and "stable" usually put out of equation.", or as I already said you wouldn't compare portage to apt, aptitude or whatever Debian/Ubuntu have to offer.
                  Testing is a topic one has to take seriously, but again I was running Gentoo for years on multiple Desktop/Laptop workstations (which are far more complicated than an appliance with fifteen packages.), and I was so impressed that I can hardly imagine better workstation system (Stable, no need to reinstall never ever.). But hey I like Ubuntu too, but neither one nor another are my favorite.
                  Last edited by reCAPTCHA; 23 October 2015, 05:52 PM.

                  Comment


                  • #59
                    Originally posted by reCAPTCHA View Post

                    Are you aware that NASDAQ uses Gentoo, so as Google (ie for Chrome OS, or that router from Phoronix news.), NASA for exploration rovers, lot of scientific projects use it, and it is really great system for appliances.
                    They're using it as a base for custom solutions they build on top of it.

                    That's a lot like saying a user running Ubuntu is running Debian because the base system is built on top of Debian. Sort of true, but not really a useful observation.

                    I would agree that Gentoo is great for that purpose, though.

                    Comment

                    Working...
                    X