Re: [PATCH V2] KVM Test: Add ioquit test case

[Lucas Meneghel Rodrigues <lmr@redhat.com>]

On Fri, 2010-05-14 at 17:43 +0800, Feng Yang wrote:
> Emulate the powercut under IO workload(dd so far) using kill -9.
> Then check image in post command.
> This case want to make sure powercut under IO workload will not
> break qcow2 image.

The big question that came to my mind here is: Are we really expected to
keep the hard disk image consistency even if the power is cut down? I am
not sure about that. Dor, what is the expected behavior on this
situation?

> Now it only work on linux.

I have a couple of minor comments on this one:

> Signed-off-by: Feng Yang <fyang@redhat.com>
> ---
>  client/tests/kvm/tests/ioquit.py       |   39 ++++++++++++++++++++++++++++++++
>  client/tests/kvm/tests_base.cfg.sample |   10 +++++++-
>  2 files changed, 48 insertions(+), 1 deletions(-)
>  create mode 100644 client/tests/kvm/tests/ioquit.py
> 
> diff --git a/client/tests/kvm/tests/ioquit.py b/client/tests/kvm/tests/ioquit.py
> new file mode 100644
> index 0000000..a202297
> --- /dev/null
> +++ b/client/tests/kvm/tests/ioquit.py
> @@ -0,0 +1,39 @@
> +import logging, time, random
> +from autotest_lib.client.common_lib import error
> +import kvm_test_utils
> +
> +
> +def run_ioquit(test, params, env):
> +    """
> +    Emulate the poweroff under IO workload(dd so far) using kill -9.
> +
> +    @param test: Kvm test object
> +    @param params: Dictionary with the test parameters.
> +    @param env: Dictionary with test environment.
> +    """
> +
> +    vm = kvm_test_utils.get_living_vm(env, params.get("main_vm"))
> +    session = kvm_test_utils.wait_for_login(vm,
> +                  timeout=int(params.get("login_timeout", 360)))
> +    session2 = kvm_test_utils.wait_for_login(vm,
> +                  timeout=int(params.get("login_timeout", 360)))
> +    try:
> +        bg_cmd = params.get("background_cmd")
> +        logging.info("Add IO workload for guest OS.")
> +        (s, o) = session.get_command_status_output(bg_cmd, timeout=60)
> +        check_cmd = params.get("check_cmd")
> +        (s, o) = session2.get_command_status_output(check_cmd, timeout=60)
> +        if int(o) <= 0:
> +            raise error.TestError("Fail to add IO workload for Guest OS")
> +
> +        logging.info("Sleep for a while")
> +        time.sleep(random.randrange(30,100))
> +        (s, o) = session2.get_command_status_output(check_cmd, timeout=300)
> +        if int(o) <= 0:
> +            logging.info("Background command finish before kill VM")

^ "IO workload finished before the VM was killed"

> +        logging.info("Kill the virtual machine")
> +        vm.process.close()
> +    finally:
> +        session.close()
> +        session2.close()
> diff --git a/client/tests/kvm/tests_base.cfg.sample b/client/tests/kvm/tests_base.cfg.sample
> index bb3646c..4387a36 100644
> --- a/client/tests/kvm/tests_base.cfg.sample
> +++ b/client/tests/kvm/tests_base.cfg.sample
> @@ -389,7 +389,11 @@ variants:
>                  rebase_mode = unsafe
>                  image_name_snapshot1 = sn1
>                  image_name_snapshot2 = sn2
> -
> +    - ioquit:
> +        type = ioquit
> +        background_cmd = "for i in 1 2 3 4; do (nohup dd if=/dev/urandom of=/tmp/file bs=102400 count=10000000 &) done"
> +        check_cmd = ps -a |grep dd |wc -l

^ Here we can add login_timeout explicitly so people can tune this
timeout.

> +        
>      # system_powerdown, system_reset and shutdown *must* be the last ones
>      # defined (in this order), since the effect of such tests can leave
>      # the VM on a bad state.
> @@ -1347,6 +1351,10 @@ variants:
>          pre_command += " scripts/hugepage.py /mnt/kvm_hugepage;"
>          extra_params += " -mem-path /mnt/kvm_hugepage"
>  
> +ioquit:
> +    post_command_noncritical = no
> +    only qcow2
> +    only Linux
>  
>  variants:
>      - @no_pci_assignable: