I've got rackspace-cloudservers-uk working for a single node cluster, but
when I ask for two worker machines, I see stack traces pointing to a lump
of JSON coming back
org.jclouds.rest.InsufficientResourcesException: {
"overLimit" : {
"code" : 413,
"message" : "OverLimit Retry...",
"details" : "Error Details...",
"retryAt" : "2012-11-14T19:34:04UTC"
}
}
This is clearly some throttling thing -either rate of RESTy requests or
rate of API calls in general -I'll have to understand that.
When the request fails, cluster launch is considered failed, so
the /Users/stevel/.whirr/hadoop/instances file doesn't get created
The machines are there though:
$ nova list
+--------------------------------------+------------+--------+-----------------------------------------------------------------------------------+
| ID | Name | Status | Networks
|
+--------------------------------------+------------+--------+-----------------------------------------------------------------------------------+
| e13cf52f-4290-4b4a-92e1-24f9922a4209 | hadoop-1f2 | ACTIVE |
public=5.79.7.206, 2a00:1a48:7803:0107:176c:89a0:ff08:2211;
private=10.177.196.41 |
| 1eb6df60-d55a-4c40-9279-f2dbd396555c | hadoop-351 | ACTIVE |
public=2a00:1a48:7803:0107:176c:89a0:ff08:220f, 5.79.7.85;
private=10.177.195.157 |
| 08bbfbcb-e37c-4335-b2d1-4d99e6083ab1 | hadoop-b19 | ACTIVE |
public=2a00:1a48:7803:0107:176c:89a0:ff08:220e, 5.79.7.90;
private=10.177.197.33 |
+--------------------------------------+------------+--------+-----------------------------------------------------------------------------------+
this means that the request did go through eventually, but wasn't picked up
-so whirr is leaking vm instances
