The main issue is the change to make the model immutable after building and it exposed an issue where large dependencyManagement sections are re-built over and over and are not cached properly.
If you want to take a look happy to point you at the code, but it will be something we'll look at in the next release or two. On Mar 23, 2015, at 8:44 AM, Jörg Schaible <joerg.schai...@swisspost.com> wrote: > Hi, > > I already reported this for 3.2.3 > (http://thread.gmane.org/gmane.comp.jakarta.turbine.maven.devel/113979), > provided Igor with an archive containing anything to run validate (I have it > still) and this resulted in https://jira.codehaus.org/browse/MNG-5669. > However, the issue in still unresolved and the situation did not improve. > > More answers inline. > > Karl Heinz Marbaise wrote: > >> Hi, >> >> have you pinned all plugins correctly and get no difference in the >> plugins version during the different executions? > > Pinned. > >> Have you defined any kind of profiles etc. or memory configurations? >> Which JDK are you using? OS is of course interesting as well... > > Gentoo Linux-3.18.7 > > Java(TM) SE Runtime Environment (build 1.7.0_76-b13) > Java HotSpot(TM) 64-Bit Server VM (build 24.76-b04, mixed mode) > >> Are you using the exact same machine? > > Yes. > >> Or running on a CI system which >> uses different nodes to execute? > > No. > >> Which filesystem is under the hood on >> this machine? > > ext4 on SSD. > > >> On 3/21/15 12:00 PM, Robert Scholte wrote: >>> It also seems to be a time-issue. >> >> looks odd... >> >> >>> I would like to see these numbers over, let's say, 10 runs. (OS and JDK >>> the same of course) >> >> +1 for this... > > The results are rather constant. In contrast to the results in my original > mail I have now 522 projects in the build. Still no problem for M305 with > 1.5GB, but M331 runs OOM with 3GB now. > > 10 times with M305 (-Xmx1536m): > ===================== %< =================== > [INFO] Total time: 21.360s > [INFO] Final Memory: 366M/1103M > [INFO] Total time: 17.612s > [INFO] Final Memory: 367M/1122M > [INFO] Total time: 20.013s > [INFO] Final Memory: 368M/1101M > [INFO] Total time: 19.400s > [INFO] Final Memory: 370M/1105M > [INFO] Total time: 16.759s > [INFO] Final Memory: 369M/1152M > [INFO] Total time: 17.997s > [INFO] Final Memory: 369M/1148M > [INFO] Total time: 18.993s > [INFO] Final Memory: 367M/1100M > [INFO] Total time: 19.385s > [INFO] Final Memory: 369M/1103M > [INFO] Total time: 17.020s > [INFO] Final Memory: 366M/1044M > [INFO] Total time: 18.984s > [INFO] Final Memory: 369M/1103M > ===================== %< =================== > > 10 times with M331 (-Xmx4g): > ===================== %< =================== > [INFO] Total time: 53.801 s > [INFO] Finished at: 2015-03-23T13:22:39+01:00 > [INFO] Final Memory: 1937M/3591M > [INFO] Total time: 50.849 s > [INFO] Finished at: 2015-03-23T13:23:35+01:00 > [INFO] Final Memory: 1940M/3641M > [INFO] Total time: 51.883 s > [INFO] Finished at: 2015-03-23T13:24:32+01:00 > [INFO] Final Memory: 1946M/3571M > [INFO] Total time: 48.187 s > [INFO] Finished at: 2015-03-23T13:25:26+01:00 > [INFO] Final Memory: 1935M/3641M > [INFO] Total time: 53.555 s > [INFO] Finished at: 2015-03-23T13:26:24+01:00 > [INFO] Final Memory: 1937M/3543M > [INFO] Total time: 51.795 s > [INFO] Finished at: 2015-03-23T13:27:21+01:00 > [INFO] Final Memory: 1949M/3569M > [INFO] Total time: 54.271 s > [INFO] Finished at: 2015-03-23T13:28:21+01:00 > [INFO] Final Memory: 1936M/3501M > [INFO] Total time: 47.280 s > [INFO] Finished at: 2015-03-23T13:29:13+01:00 > [INFO] Final Memory: 1947M/3601M > [INFO] Total time: 47.663 s > [INFO] Finished at: 2015-03-23T13:30:06+01:00 > [INFO] Final Memory: 1945M/3527M > [INFO] Total time: 55.549 s > [INFO] Finished at: 2015-03-23T13:31:07+01:00 > [INFO] Final Memory: 1935M/3502M > ===================== %< =================== > >>> One thing worth investigating is if the buildplan related instances can >>> be cleaned up. > > The build only contains what's currently required. However, in case of HEAD > it is nearly anything. > >>> >>> thanks, >>> Robert >>> >>> Op Wed, 18 Mar 2015 11:19:33 +0100 schreef Jörg Schaible >>> <joerg.schai...@swisspost.com>: >>> >>>> Hi, >>>> >>>> it seems to get my standard complaint about every new Maven version, but >>>> we're currently stuck to Maven 3.0.5. One reason is the vast memory >>>> usage of >>>> any later version. >> >> Ok you have one reason to stuck with Maven 3.0.5...what are the other >> reasons ? > > In case of a branch we have a lot less projects in the build, but in that > case M3 fails for us to build them in proper order and we even have to use > M221. Different story though (MNG-5207). > > We're also stuck with the site plugin version 3.0-beta-3 (resp. 2.2 for > M221) because of MSITE-604 and MSITE-664. > > [snip] > > Cheers, > Jörg > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org > For additional commands, e-mail: dev-h...@maven.apache.org > Thanks, Jason ---------------------------------------------------------- Jason van Zyl Founder, Takari and Apache Maven http://twitter.com/jvanzyl http://twitter.com/takari_io --------------------------------------------------------- Our achievements speak for themselves. What we have to keep track of are our failures, discouragements and doubts. We tend to forget the past difficulties, the many false starts, and the painful groping. We see our past achievements as the end result of a clean forward thrust, and our present difficulties as signs of decline and decay. -- Eric Hoffer, Reflections on the Human Condition --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
