[one-users] One step at a time, now there are SCP issues (Shell Initilization problem?)

[Daniel Molina]

Hi,

On 31 July 2012 05:43, Jason Fairfield <pegasusjf at siu.edu> wrote:
> This is technically a sequel for this post:
> http://lists.opennebula.org/pipermail/users-opennebula.org/2012-July/009698.html.
> However, I made some progress and encounted a new error with scp.  I listed
> my changes since that last post (adding nfs) to double check, so please bear
> with me, first time doing this.  Plus some unrelated quick questions at the
> end.
>
> Using XUbuntu (12.04) for the front end and Debian (latest) on the host
>
> It seems while SSH works fine, scp doesn't.  I failed the test listed at
> http://www.openssh.org/faq.html#2.9.  (At least at first, it asked for key
> authentication, which is exited out of, now it says file not found)
>
> It says change my shell to get rid of output for non-interactive sessions.
> How do I do that?  The Debian .bashrc file (the only one, with .profile) in
> that list has nothing that seems to spit out the generic stuff that gets
> returned when you
>
> Now, the changes:
>
> I created /srv/cloud/one01/var/datastores/0 in the host (I even tried adding
> 8 to 0).  I double checked the permissions (all oneadmin 755).
>
> I set onedatastore update to this:
>
> DS_MAD="fs"
> TM_MAD="ssh"

If you set TM_MAD to "ssh", nfs is not required anymore and the images
will be copied to the remote hosts instead of using the nfs server.

>
> and I added to /etc/exports (I overlooked this part)
> /srv/cloud/one01/var host09(ro,sync,no_subtree_check,no_root_squash)
>
>
> and to /etc/fstab on host09 in accordance to the open nebula 3 book
>
> on-front:/srv/cloud/one01/var /srv/cloud/one01/var
>  nfs udp,_netdev 0 0
>
>
>
> These seem to be the pertinent logs in oned.log.
>
> Mon Jul 30 21:34:44 2012 [DiM][D]: Deploying VM 10
> Mon Jul 30 21:34:44 2012 [TM][D]: Message received: LOG I 10 Command
> execution fail: /srv/cloud/one01/var/remotes/tm/ssh/clone
> cloudadmin-OptiPlex-380:/srv/cloud/one01/var/datastores/1/44eb8aeca94df8d12223fac727782957
> host09:/srv/cloud/one01/var//datastores/0/10/disk.0 10 1
>
> Mon Jul 30 21:34:44 2012 [TM][D]: Message received: LOG I 10 clone: Cloning
> cloudadmin-OptiPlex-380:/srv/cloud/one01/var/datastores/1/44eb8aeca94df8d12223fac727782957
> in /srv/cloud/one01/var/datastores/0/10/disk.0
>
> Mon Jul 30 21:34:44 2012 [TM][D]: Message received: LOG E 10 clone: Command
> "scp -r
> cloudadmin-OptiPlex-380:/srv/cloud/one01/var/datastores/1/44eb8aeca94df8d12223fac727782957
> host09:/srv/cloud/one01/var//datastores/0/10/disk.0" failed: ssh: connect to
> host cloudadmin-OptiPlex-380 port 22: Connection refused
>
> Mon Jul 30 21:34:44 2012 [TM][D]: Message received: LOG E 10 Error copying
> cloudadmin-OptiPlex-380:/srv/cloud/one01/var/datastores/1/44eb8aeca94df8d12223fac727782957
> to host09:/srv/cloud/one01/var//datastores/0/10/disk.0
>
> Mon Jul 30 21:34:44 2012 [TM][D]: Message received: LOG I 10 ExitCode: 1
>
> Mon Jul 30 21:34:44 2012 [TM][D]: Message received: TRANSFER FAILURE 10
> Error copying
> cloudadmin-OptiPlex-380:/srv/cloud/one01/var/datastores/1/44eb8aeca94df8d12223fac727782957
> to host09:/srv/cloud/one01/var//datastores/0/10/disk.0
>
> Mon Jul 30 21:34:45 2012 [ReM][D]: VirtualMachineAction method invoked
>

Did you check the installation OpenNebula guide? It looks like you do
not have ssh access to the target host, did you include the oneadmin
public key in the target host? or maybe iptables enabled?

> Here are the small_network.net, image.one, and ttylinux.one (doesn't like
> using the names) respectively
>
> NAME = "Small network"
> TYPE = FIXED
>
> #Now we'll use the cluster private network (physical)
> BRIDGE = lan0
> LEASES = [ IP="192.168.0.5"]
> LEASES = [ IP="192.168.0.6"]
> LEASES = [ IP="192.168.0.7"]
>
>
> image.one (I tried the http version, that didn't work either)
>
> NAME = ttylinux
> PATH = "/tmp/ttylinux/ttylinux.img"
> TYPE = OS
>
> and ttylinux.one (it doesn't like using the names)
>
> NAME   = ttylinux
> CPU    = 0.1
> MEMORY = 64
>
> DISK   = [ IMAGE_ID = 6 ]
> NIC    = [ NETWORK_ID = 6 ]
> FEATURES=[ acpi="no" ]
>
> quick questions:
>
> How do I reply to replies (thank you by the way! :) ) and have it show on
> the list?  cc to users at lists.opennebula.org

You just have to reply the mail and include users at lists.opennebula.org in cc.

>
> I've deleted and remade images, vnets, and vms, the ID numbers keep
> climbing, how do I reset the count?

OpenNebula will keep any resource in its DB for accounting, so the
only way to reset the counters is deleting the DB.

>
> Is nfs absolutely required?  One thing I missed in my previous try was I
> didn't install the nfs-kernel-server and configure on both machines.  Just
> need a quick answer.
>

If you use TM_MAD="nfs", it is required.

Hope this helps.

-- 
Daniel Molina
Project Engineer
OpenNebula - The Open Source Solution for Data Center Virtualization
www.OpenNebula.org | dmolina at opennebula.org | @OpenNebula