Re: [PATCH] KVM Test: add -w parameter in nc command in kvm_utils.py

[Michael Goldish <mgoldish@redhat.com>]

On 10/21/2010 03:57 AM, Feng Yang wrote:
> 
> ----- "Michael Goldish" <mgoldish@redhat.com> wrote:
> 
>> From: "Michael Goldish" <mgoldish@redhat.com>
>> To: "Feng Yang" <fyang@redhat.com>
>> Cc: autotest@test.kernel.org, kvm@vger.kernel.org
>> Sent: Wednesday, October 20, 2010 6:48:42 PM GMT +08:00 Beijing / Chongqing / Hong Kong / Urumqi
>> Subject: Re: [PATCH] KVM Test: add -w parameter in nc command in kvm_utils.py
>>
>> On 10/20/2010 12:18 PM, Feng Yang wrote:
>>>
>>> ----- "Michael Goldish" <mgoldish@redhat.com> wrote:
>>>
>>>> From: "Michael Goldish" <mgoldish@redhat.com>
>>>> To: "Feng Yang" <fyang@redhat.com>
>>>> Cc: autotest@test.kernel.org, kvm@vger.kernel.org
>>>> Sent: Wednesday, October 20, 2010 5:11:32 PM GMT +08:00 Beijing /
>> Chongqing / Hong Kong / Urumqi
>>>> Subject: Re: [PATCH] KVM Test: add -w parameter in nc command in
>> kvm_utils.py
>>>>
>>>> On 10/20/2010 08:55 AM, Feng Yang wrote:
>>>>> If a connection and stdin are idle for more than timeout seconds,
>>>>> then the connection is silently closed. withou this paramter,
>>>>> nc will listen forever for a connection. This may cause our test
>>>>> hang. redmine issue:
>>>>> http://redmine.englab.nay.redhat.com/issues/show/6947
>>>>> Add -w parameter should fix this issue.
>>>>>
>>>>> Signed-off-by: Feng Yang <fyang@redhat.com>
>>>>> ---
>>>>>  client/tests/kvm/kvm_utils.py |    2 +-
>>>>>  1 files changed, 1 insertions(+), 1 deletions(-)
>>>>>
>>>>> diff --git a/client/tests/kvm/kvm_utils.py
>>>> b/client/tests/kvm/kvm_utils.py
>>>>> index a51c857..526d38d 100644
>>>>> --- a/client/tests/kvm/kvm_utils.py
>>>>> +++ b/client/tests/kvm/kvm_utils.py
>>>>> @@ -733,7 +733,7 @@ def remote_login(client, host, port,
>> username,
>>>> password, prompt, linesep="\n",
>>>>>      elif client == "telnet":
>>>>>          cmd = "telnet -l %s %s %s" % (username, host, port)
>>>>>      elif client == "nc":
>>>>> -        cmd = "nc %s %s" % (host, port)
>>>>> +        cmd = "nc %s %s -w %s" % (host, port, timeout)
>>>>>      else:
>>>>>          logging.error("Unknown remote shell client: %s" %
>> client)
>>>>>          return
>>>>
>>>> I don't understand how remote_login() can stall here.
>>>> kvm_utils._remote_login() doesn't rely on nc's self-termination. 
>> If
>>>> no
>>>> shell prompt is available after 10 seconds, nc is forcefully
>>>> terminated
>>>> and _remote_login() fails.  If it somehow stalls this might
>> indicate
>>>> a
>>>> bug somewhere.
>>> It is really rarely reproduce. Only meet this issue when guest panic
>> or core dump.
>>> nc stall. Do not know why it is not terminated after 10s.
>>>
>>> I think -w parameter is helpful in this situation.
>>> If it does not cause other issue,  we'd better add -w in nc
>> command.
>>>
>>> Thanks for your command. What do you think?
>>>
>>> Feng Yang
>>
>> Adding -w 10 may cause trouble.  If a good functional session is idle
>> for more than 10 seconds during a test, it will be closed.  For
>> example,
>> if you run a command that takes more than 10 seconds to complete, and
>> produces no output while it runs, the session will be closed.
> Seems the session also will be closed in our current code in this situation.
> If no output for 10s, _remote_login() will return False. Then session will be closed in remote_login().

If the login process takes more than 10s, yes, the session will be
closed.  However, if we use -w 10, the session will be closed whenever
it becomes idle, not just during login.  So if we login successfully and
then run a test (e.g. an autoit script), after 10s of no output the test
will be terminated.

> I have run some cases, it works well.  But I think we'd better first post this code, in case it really cause trouble.
> 
> I will do more test on it.
> 
>>
>> Also, I doubt it'll solve the problem you've experienced.  As it is,
>> the
>> code should (and usually does) properly handle the case of nc not
>> terminating.
>>
>> If you manage to reproduce this again please save the log so I can
>> have
>> a look at it.
> 
> When meet this issue again, I will send log to you. Thanks for your help!
> 
> Feng Yang
> 
> 
> 
>> --
>> To unsubscribe from this list: send the line "unsubscribe kvm" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html