Several VMs are not accessible after their sliver is shown as ready

[lnevers@bbn.com]

Set up 10 experiments but some of the hosts are not accessible even though the slivers status reports that the slivers are ready.

Scenario:

• 10 slices created, each slice contains one sliver with 2 VMs.
• Slice names are 2vmslice1 through 2vmslice10.
• All slivers are ready according to the sliver status.

There are 3 VMs on which I am prompted for a password, below are the sliver name, the hostname and ports for the VMs not allowing login.

SliceName: 2vmslice4  'hostname': 'pc3.utah.geniracks.net', 'port': 31803, 
SliceName: 2vmslice8  'hostname': 'pc1.utah.geniracks.net', 'port': 32826, 
SliceName: 2vmslice9  'hostname': 'pc1.utah.geniracks.net', 'port': 33082, 

Attaching the sliver status to the ticket.

[lnevers@bbn.com]

On 7/19/12 2:26 PM, Leigh Stoller wrote:

There are 3 VMs on which I am prompted for a password, below are the

sliver name, the hostname and ports for the VMs not allowing login.

There was a transient, unexplainable error. The worst kind. I logged in and reran the account setup script, and it worked fine. There was no error message from the first run, just a non-zero exit status.

So this remains open until I (or someone else) has time to sink into it.

Lbs

Can I delete the slivers?

Also, if I run into this again, I will let you know.

[lnevers@bbn.com]

Significant improvements have been made in this area. Numerous experiments have been run with multiple VMs, none of which showed the node access failures encountered in ticket. Closing ticket, no longer an issue.