
Dan Smith wrote:
KR> 6 - 7 iterations - seems pretty consistent in my testing.
Okay.
KR> In the future, most of these tests will use the provider to create the KR> guest (not virsh), so I think this problem will go away on its own in KR> time.
Indeed. As long as that's the plan, I think doing this for those tests that need to check this behavior is fine.
What's up with the name though? The function has 'virsh' in the name, but it's querying the providers, right?
It's checking both virsh and the providers. The hack here is that virsh doesn't hit the caching issue. So if virsh lists the guest, then we poll until the providers see it. If virsh doesn't see the guest, then the guest wasn't created. So we don't waste time polling for a guest that doesn't exist. Even still, the name is pretty poor. I'll think something else up and resend. =) -- Kaitlin Rupert IBM Linux Technology Center kaitlin@linux.vnet.ibm.com