This issue is fixed for me. Leaving the below solution here for others. =====
Step A: Comment out "Install shellcheck" section that installs Cabal in below file. /dev-support/docker/Dockerfile The hadoop env setup completed successfully though it complained about low memory (2 GB min required) Step B: So uncomment above lines and instead increased the memory allocated to the vm via Vagrant config. 1. Add below in Vagrantfile config.vm.provider "virtualbox" do |vb| # Customize the amount of memory on the VM: 3 GB vb.memory = "3072" end 2. ~$ vagrant reload 3. ~$ vagrant ssh 4. In the hadoop folder inside the ubuntu, type ~$ ./start-build-env.sh This should now create the hadoop build env successfully without any error and the overhead is ~ 1GB memory leaving ~2GB free. Thanks, Arun On Wed, Mar 1, 2017 at 10:04 PM, Arun S. Kumar <arunsk....@gmail.com> wrote: > Hi, > > I am trying to setup Hadoop locally by following https://wiki.apache. > org/hadoop/HowToContribute > and https://git-wip-us.apache.org/repos/asf?p=hadoop.git;a= > blob;f=BUILDING.txt. > > I am on MacOSX Sierra (10.12.3) 8GB RAM 2.2GHZ intel core i7. > I installed ubuntu-trusty-64 via vagrant and virtualbox provider. > > When I ssh'ed into ubuntu and tried running start-build-env.sh , it failed > at > > Step 21/33 : RUN cabal update > > Killed > > The command '/bin/sh -c cabal update' returned a non-zero code: 137 > I googled around and found one possible reason could be that it doesn't > have enough memory > > while creating the container and tried a suggestion to restart VM but > again got different error > below. > > Step 21/33 : RUN cabal update > > cabal: out of memory (requested 1048576 bytes) > > The command '/bin/sh -c cabal update' returned a non-zero code: 1 > Is Cabal really needed or optional ? Any help is much appreciated. > > TIA. > > --------------------------------- > Regards, > Arun S. Kumar > > -- --------------------------------- Regards, Arun S. Kumar