Re: [Qemu-devel] QMP: RFC: I/O error info & query-stop-reason

[Anthony Liguori <anthony@codemonkey.ws>]

On 06/03/2011 08:44 AM, Luiz Capitulino wrote:
> On Fri, 3 Jun 2011 14:39:41 +0100
> "Daniel P. Berrange"<berrange@redhat.com>  wrote:
>
>> On Fri, Jun 03, 2011 at 08:26:56AM -0500, Anthony Liguori wrote:
>>> On 06/03/2011 07:57 AM, Daniel P. Berrange wrote:
>>>> On Fri, Jun 03, 2011 at 07:43:24AM -0500, Anthony Liguori wrote:
>>>>> On 06/03/2011 04:26 AM, Daniel P. Berrange wrote:
>>>>> errors stop a guest) instead of trying to model an internal QEMU
>>>>> concept (vm_stop()).
>>>>>
>>>>> If you have other user visible concepts that you want to know about,
>>>>> please share the use-cases and we can think about how to model it
>>>>> such that it's not exposing internal QEMU details.
>>>>
>>>> None of the requested info is exposing internal QEMU impl details
>>>> with one exception. The reasons are either administrative commands,
>>>> host OS failures, guest OS failures, or the exception, KVM internal
>>>> emulation failure.
>>>>
>>>> The core problem is that an app connects to QEMU, finds it is paused,
>>>> and wants to decide what action to take. If the guest is paused due
>>>> to a previous admin 'stop' command,
>>>
>>> Let's be very clear here.  QEMU does not provide a way to figure out
>>> what the previous QMP user did.  That is not a use case we support
>>> today and it's not one we can support by just adding a reason to
>>> stop.  It's far more complicated than just that.
>>>
>>>> it will allow resuming. If it is
>>>> paused due to guest OS poweroff,
>>>
>>> This is legitimate but only occurs if you use -no-shutdown.  So
>>> having a query-state have a "powered-off" flag would be a Good
>>> Thing.
>>>
>>>> it might decide to issue a 'system_reset'
>>>> command and then 'resume'. If it is paused due to watchdog,
>>>
>>> I think what we're getting at is the need for an enumeration.  So
>>> let's introduce one.  Here's what I propose:
>>>
>>> SQMP
>>> query-status
>>> ------------
>>>
>>> Return a json-object with the following information:
>>>
>>> - "running": true if the VM is running, or false if it is paused (json-bool)
>>> - "singlestep": true if the VM is in single step mode,
>>>                  false otherwise (json-bool)
>>> - "status": one of the following values (json-string) (optional)
>>>        "prelaunch" - QEMU was started with -S and guest has not started
>>>        "running" - guest is actively running
>>>        "singlestep" - guest is running in single step mode
>>>        "paused" - guest has been paused via the 'stop' command
>>>        "postmigrate" - guest is paused following a successful 'migrate'
>>>        "shutdown" - guest is shut down (and -no-shutdown is in use)
>>>        "io-error" - the last IOP has failed and the device is
>>> configured to pause on I/O errors
>>>        "watchdog-error" - the watchdog action is configured to pause
>>> and has been triggered
>>
>> Perhaps I didn't communicate well, but this pretty much matches
>> what I was trying to ask for in my previous message, so gets
>> my vote!
>
> Mine too, I like it. Expect patches next week :)
>
> My only comment is that, in case this an improved version of
> query-status we could have a new command (like query-statys2 or
> query-vm-status).

No, let's just keep query-status and add the new field.

For compatibility, you need to fall back to using running/singlestep anyway.

Regards,

Anthony Liguori

>
>