Hi Austin,
Yes the VolumeCreator class shuts down all instances in the _at_sc-volumecreator security group after successful creation of a volume by default. If you run two volume creator sessions at the same time they will attempt to use the same host. Whichever finishes first will shutdown the instance (unless you tell it not to) which breaks things for the other session. If you want to run two volume creator sessions at the same time you should pass the --no-shutdown option to the createvolume command or use shutdown_instance=False in the VolumeCreator class's constructor if you're using the class directly. After the volumes have been created successfully you can run the following command to shut down the volume host instance(s):
$ starcluster stop volumecreator
Hope that helps,
~Justin
________________________________________
From: starcluster-bounces_at_MIT.EDU [starcluster-bounces_at_MIT.EDU] On Behalf Of Austin Godber [godber_at_uberhip.com]
Sent: Sunday, August 22, 2010 7:55 PM
To: starcluster_at_mit.edu
Subject: [Starcluster] InvalidInstanceID.NotFound on Disk Creation
Has anyone else seen an InvalidInstanceID error when creating a disk?
I got the following error when creating a disk:
volume.py:186 - ERROR - exception thrown: EC2ResponseError: 400 Bad Request
<?xml version="1.0"?>
<Response><Errors><Error><Code>InvalidInstanceID.NotFound</Code><Message>The
instance ID 'i-913c54fb' does not
exist</Message></Error></Errors><RequestID>XXXXXX</RequestID></Response>
The instance definitely started since I see it as terminated in my AWS
console. Oh, I bet I know what happened. I was actually creating TWO
disks at the same time. I bet that disk creation shuts down all
instances in the _at_sc-volumecreator group. And when the first one was
killed, the second one was killed along with it.
Austin
_______________________________________________
Starcluster mailing list
Starcluster_at_mit.edu
http://mailman.mit.edu/mailman/listinfo/starcluster
Received on Sun Aug 22 2010 - 22:06:52 EDT