Looks like there's been quite a bit of work on both the Capacity Scheduler and the Fair Scheduler. Are there others we should look at in 20.0?
Again, thank you for your informative responses. David Raghu Angadi wrote: > > I think 0.20.0 is preferred because of new job scheduler(s) there. > Stability wise, 0.20 is expected to be as good as 0.19.2. > > I don't know the schedule for 0.19.2. > > Raghu. > > David Ritch wrote: >> Raghu, >> >> Thank you for your prompt and informative response. Moving to >> anything that >> ends in .0 is a bit scary - what are the reasons to go with 0.20.0 >> instead >> of 0.19.2? Yahoo is jumping from 0.18.x directly to 0.20.0? Why is >> Yahoo >> skipping the 0.19.x release? >> >> Is the expectation that 0.19.2 will be released at the same time as >> 0.20.0? >> >> Thanks, >> >> David >> >> On Wed, Mar 18, 2009 at 1:31 PM, Raghu Angadi <rang...@yahoo-inc.com> >> wrote: >> >>> Short is answer I am afraid is no. >>> >>> As an alternative, I recommend upgrading to latest 0.19.x or 0.20.0 >>> (to be >>> released in couple of days). 0.19.2 is certainly a lot better than >>> 0.19.0. >>> Yahoo is rolling out 0.20.x if that helps your confidence. >>> >>> Raghu. >>> >>> >>> David Ritch wrote: >>> >>>> There is an established procedure for upgrading from one release of >>>> Hadoop >>>> to a newer release. Is there something similar to move back to an >>>> lower-numered release? >>>> >>>> Specifically, we have data in a cloud running Hadoop-19.0. Because of >>>> stability issues, we are wondering whether we should move back to >>>> 18, but >>>> we >>>> don't want to lose our data. Is there a downward migration path? >>>> >>>> Thanks, >>>> >>>> David >>>> >>>> >> > >
