greese / dasein-cloud-cloudstack Goto Github PK

There's a way to launch a VM which will fail to launch due to insufficient capacity, the VM will remain in the CS inventory in an error state, however Dasein will not return it in its listVirtualMachines.

https://github.com/greese/dasein-cloud-cloudstack/blob/master/src/main/java/org/dasein/cloud/cloudstack/compute/VirtualMachines.java#L697

As far as I can see, this cache will never expire, which means that once we have a list of products we don't pick up changed values or additional items without restarting the component that has dasein memory-resident.

I'm thinking a 1-2 hour timeout seems appropriate... certainly doesn't need to be called often, so it can be fairly long.

There is an infinite loop inside the waitForJob method of CSCloud that needs a timeout.

(public Document waitForJob(String, String))

Binod

The driver is trying to set startPort and endPort where it should be setting icmpType and icmpCode.