readme instructing not correct

[ghost]

on windows

i booted the 4 nodes.
had to run it a few times for it to complete.

When sshing into master, the only sh file in the root directory is "postinstall.sh", not "prepare-cluster.sh".
Also start-all.sh is not there.

Would appreciate confirmation o this so i know i am doing things correctly...

[fs111]

Can you post all the steps you did and also all their output, if you still have it?

Also, listing your windows/vbox/vagrant versions will be helpful.

Thanks!

[ghost]

thanks for the quick feedback... thats was fast

i dont have all the output logged. If you can tell me how to do that i am happy to run it all again.

here is what i did:

1. "vagrant up"
   This took a few tries. I looked in the virtual box GUI to see the boxes being made.
   I just reran "vagrant up" 4 times, and it eventually completed.
2. vagrant ssh master
   "whoami", tells me i am "vagrant"user.
   ls, and i could only see postinstall.sh, so i did "sudo ./postinstall.sh"

This ran to completion. lots of apt-get stuff.
If you can tell me how to get the logs of it, I can add it to this issue.

cheers

[ghost]

I am using the latest version of vagrant (1.6.5) & virtual box (4.3.16)

[fs111]

Please leave the postinstall script alone. It is an artifact of the vagrant base box and has nothing to do with our setup.

Since you mentioned, that you had a few problems getting the boxes up, I guess they are not fully provisioned yet. Please run "vagrant provision" to do that.

Also, where in the world are you located? We sometimes run into time-outs if we can't find a fast apache mirror close to you.

[ghost]

Ok. I have already run it :) hehe
I should start from scratch ?

i am in Berlin, Germany.

Here is the output from running "vagrant provision"
C:_data_temp\vagrant-cascading-hadoop-cluster>vagrant provision
==> hadoop1: Running provisioner: puppet...
Shared folders that Puppet requires are missing on the virtual machine.
This is usually due to configuration changing after already booting the
machine. The fix is to run a vagrant reload so that the proper shared
folders will be prepared and mounted on the VM.

C:_data_temp\vagrant-cascading-hadoop-cluster>

[fs111]

please do a "vagrant reload --provision", that should bring things back to a consistent state.

Since I am also in Berlin, I think the downloads will work from here...

[ghost]

small world....

i shut down all the vm's
i ran vagrant up (this did its thing), then the "vagrant reload --provision"

i noticed in the log that the version of guest additions on my host, and in the vm's is different.
is this the problem ?
Also, the command stopped at node hadoop1 (i waited 10 minutes)

LOG BELOW:

C:_data_temp\vagrant-cascading-hadoop-cluster>vagrant reload --provision
==> hadoop1: Attempting graceful shutdown of VM...
==> hadoop1: Clearing any previously set forwarded ports...
==> hadoop1: Clearing any previously set network interfaces...
==> hadoop1: Preparing network interfaces based on configuration...
hadoop1: Adapter 1: nat
hadoop1: Adapter 2: hostonly
==> hadoop1: Forwarding ports...
hadoop1: 22 => 2222 (adapter 1)
==> hadoop1: Running 'pre-boot' VM customizations...
==> hadoop1: Booting VM...
==> hadoop1: Waiting for machine to boot. This may take a few minutes...
hadoop1: SSH address: 127.0.0.1:2222
hadoop1: SSH username: vagrant
hadoop1: SSH auth method: private key
hadoop1: Warning: Connection timeout. Retrying...
==> hadoop1: Machine booted and ready!
==> hadoop1: Checking for guest additions in VM...
hadoop1: The guest additions on this VM do not match the installed version of
hadoop1: VirtualBox! In most cases this is fine, but in rare cases it can
hadoop1: prevent things such as shared folders from working properly. If you see
hadoop1: shared folder errors, please make sure the guest additions within the
hadoop1: virtual machine match the version of VirtualBox you have installed on
hadoop1: your host and reload your VM.
hadoop1:
hadoop1: Guest Additions Version: 4.2.0
hadoop1: VirtualBox Version: 4.3
==> hadoop1: Setting hostname...
==> hadoop1: Configuring and enabling network interfaces...
==> hadoop1: Mounting shared folders...
hadoop1: /vagrant => C:/_data/_temp/vagrant-cascading-hadoop-cluster
hadoop1: /tmp/vagrant-puppet-1/manifests => C:/_data/_temp/vagrant-cascading-hadoop-cluster/mani
fests
hadoop1: /tmp/vagrant-puppet-1/modules-0 => C:/_data/_temp/vagrant-cascading-hadoop-cluster/modu
les
==> hadoop1: Running provisioner: puppet...
==> hadoop1: Running Puppet with datanode.pp...
==> hadoop1: stdin: is not a tty
==> hadoop1: notice: /Stage[main]/Base/Exec[apt-get update]/returns: executed successfully

[fs111]

This is odd. It could be one of the famous windows vs. unix line ending problems... How did you clone the project? How is your git configured w.r.t. line endings?

[ghost]

ok makes sense.

i will set git line ending globally, and then start fresh with a new git pull.

will take a few minutes :)

[ghost]

BTW used "git config --global core.autocrlf true"

[fs111]

so, did it work now?

[ghost]

i did not get a chance. I will do it on Ubuntu instead. Must put this on back burner for a bit though; got to work on GUI

[esoterick]

So as an easy "Fix" just download the repo as a zip and everything will work until stuff is patched.