Re: [Libvirt-cim] [PATCH] [TEST] #2 Fix CS 22 to use providers instead of virsh
>
> -#define VS
> + status, test_network = create_netpool_conf(options.ip, options.virt)
> + if status != PASS:
> + destroy_netpool(options.ip, options.virt, test_network)
>
Why do we require to call destroy_netpool() when it is not created
successfully?
Nope, this is a good point. I'll remove it and send a new patch.
> -#Suspend the defined VS
> - - try:
> - ret = cxml.run_virsh_cmd(options.ip, "suspend")
> - if not ret :
> - Globals.logger.info("Suspending defined VS %s failed, as
> expected" \
> -% test_dom)
> - status = PASS
> + status = chk_state(test_dom, options.ip, DEFINE_STATE,
> options.virt)
> + if status != PASS:
> + logger.error("%s not in defined state as expected." %
> test_dom)
> + status = FAIL +
I think the log stmt here is misleading.
We are trying to move from Define to Suspend, aint?
Correct. However, suspending a defined guest is not supported, so we
expect the call to RSC() to fail. Because of that, we expect the guest
to remain in the defined state. If the guest is in some other state,
then we've encountered some unknown behavior and should fail.
I think the log stmt should be as below:
logger.error("%s not in suspend state as expected." % test_dom)
The chk_state() call is verifying that the guest is in the defined
state, not the suspended state.
--
Kaitlin Rupert
IBM Linux Technology Center
kaitlin(a)linux.vnet.ibm.com