Re: StarCluster not mounting EBS volume correctly

From: Dan Tenenbaum <no email>
Date: Thu, 28 Oct 2010 12:24:41 -0700

On Wed, Oct 27, 2010 at 7:29 PM, Justin Riley <jtriley_at_mit.edu> wrote:

> On 10/27/10 9:23 PM, Dan Tenenbaum wrote:
>
>> I didn't pay close enough attention, but I think sometimes perhaps the
>> stop cluster command displays instance IDs of instances that have already
>> stopped.
>>
> Ahh, yes the stop command is picking up terminated instances because the
> security group is being destroyed/created repeatedly. This should be fixed
> in the latest github code, however, terminating already terminated instances
> is harmless, albeit obviously confusing for users.
>
>
Thanks! It was confusing. ;)


> I still need to investigate why you're getting the "failed to authenticate
> to host root as user " errors. I'll follow up with this when I get a chance
> to test.
>
>
I might know why this was happening.... The next time I tried to ssh into an
ec2 instance, I got the "SOMEONE might be doing something nasty" error.
Often this happens when EC2 recycles IP addresses and what is in my
.known_hosts file doesn't match the new key. I just delete the offending
line and I'm good to go again. This may have been the issue yesterday.



> Is anyone else experiencing this issue with createvolume in SC 0.91 (or
> 0.91.1)?
>
>

I guess you meant to send this to the list but it looks like it just came to
me. ;)

> ~Justin
>
> Dan
Received on Thu Oct 28 2010 - 15:24:44 EDT