Please see my suggestion in HAMA-395 https://issues.apache.org/jira/browse/HAMA-395?focusedCommentId=13052526#comment-13052526
2011/6/21 Edward J. Yoon <[email protected]> > Let's commit it. > > On Tue, Jun 21, 2011 at 9:38 PM, Thomas Jungblut > <[email protected]> wrote: > > Oh crap :D > > Thanks Edward, I wasn't aware of this "trick". > > > > Thank you very much! > > > > > > > > 2011/6/21 Edward J. Yoon <[email protected]> > > > >> Hi, > >> > >> I talked w/ thomas about HAMA-395 on twitter, > >> > >> Thomas Jungblut: > >> @eddieyoon the main problem is that I can't figure it out why. The BSP > >> method finishes and written out the complete results. But no cleanup > >> 13 minutes ago in reply to eddieyoon from web > >> > >> And, I'm reply to hama-dev@ so that we can share about this. > >> > >> As mentioned > >> > http://incubator.apache.org/hama/docs/r0.2.0/ApacheHama-0.2_UserGuide.pdf, > >> The BSP job will be finished only when all the processes have no more > >> local and outgoing queue entries and all processes are done, or killed > >> by the user. > >> > >> Hence, you have to clear all queues manually after escape from main > while > >> loop. > >> > >> + // Clears all queues entries. > >> + peer.clear(); > >> + // finally save the chunk of pageranks > >> + PageRankBase.savePageRankMap(peer, conf, lastTentativePagerank); > >> + LOG.info("Finished with iteration " + iteration + "!"); > >> > >> HAMA-395-v9.patch fixes that problem, and works well on my test cluster. > :) > >> > >> -- > >> Best Regards, Edward J. Yoon > >> @eddieyoon > >> > > > > > > > > -- > > Thomas Jungblut > > Berlin > > > > mobile: 0170-3081070 > > > > business: [email protected] > > private: [email protected] > > > > > > -- > Best Regards, Edward J. Yoon > @eddieyoon > -- Thomas Jungblut Berlin mobile: 0170-3081070 business: [email protected] private: [email protected]
