[libvirt] [RFC PATCH v2 1/3] qemu: Expose additional timing metrics for 'setup' and 'mbps'
Michael R. Hines
mrhines at linux.vnet.ibm.com
Fri Apr 4 05:41:21 UTC 2014
On 02/03/2014 08:52 PM, Daniel P. Berrange wrote:
> On Mon, Feb 03, 2014 at 01:32:59PM +0100, Jiri Denemark wrote:
>> On Mon, Jan 13, 2014 at 14:28:10 +0800, mrhines at linux.vnet.ibm.com wrote:
>>> From: "Michael R. Hines" <mrhines at us.ibm.com>
>>>
>>> RDMA migration uses the 'setup' state in QEMU to optionally lock
>>> all memory before the migration starts. The total time spent in
>>> this state is already exposed by QEMU, so expose it in libvirt.
>>>
>>> Additionally, QEMU also now exports migration throughput (mbps),
>>> so let's see that one too...
>>>
>>> Signed-off-by: Michael R. Hines <mrhines at us.ibm.com>
>>> ---
>>> include/libvirt/libvirt.h.in | 15 +++++++++++++++
>>> src/qemu/qemu_driver.c | 14 ++++++++++++++
>>> src/qemu/qemu_monitor.h | 12 ++++++++++++
>>> src/qemu/qemu_monitor_json.c | 8 ++++++++
>>> 4 files changed, 49 insertions(+)
>>>
>>> diff --git a/include/libvirt/libvirt.h.in b/include/libvirt/libvirt.h.in
>>> index 5aad75c..5ac2694 100644
>>> --- a/include/libvirt/libvirt.h.in
>>> +++ b/include/libvirt/libvirt.h.in
>>> @@ -4153,6 +4153,21 @@ int virDomainAbortJob(virDomainPtr dom);
>>> #define VIR_DOMAIN_JOB_DOWNTIME "downtime"
>>>
>>> /**
>>> + * VIR_DOMAIN_JOB_MBPS:
>>> + *
>>> + * virDomainGetJobStats field: network throughput used while migrating.
>>> + */
>>> +#define VIR_DOMAIN_JOB_MBPS "mbps"
>> I think this would be better as
>>
>> #define VIR_DOMAIN_JOB_THROUGHPUT "throughput"
>>
>> And you need to explicitly mention the type of the value returned in
>> this parameter.
> Hmm, in block I/O tuning XML/APIs we use either
>
> VIR_DOMAIN_BLOCK_IOTUNE_TOTAL_BYTES_SEC
>
> or
>
> VIR_DOMAIN_BLKIO_DEVICE_READ_BPS
>
> which would suggest that first we should measure this in bps too,
> not mbp, and using either BPS or BYTES_SEC as a constant suffix
> is reasonable for consistency.
>
> Daniel
Acknowledged.
- Michael
More information about the libvir-list
mailing list