Re: [pcp] Difference between: kernel.percpu.cpu.user and kernel.pernode.cpu.user

[Mark Goodwin <mgoodwin@xxxxxxxxxx>]

On 11/27/2014 07:34 PM, Shirshendu Chakrabarti wrote:
[...]
> I am running on Amazon Linux - 3.4.71-63.amzn1.x86_64
is that an SMP kernel?

> I am running, PCP-3-10.0-1.
>
> Please see its a t1.micro instance.
are there any 'cpu*' symlinks below /sys/devices/system/node/node0 ?
and does /sys/devices/system/node/node0/cpumap exist?  This is where
the PCP "linux" PMDA gets it's node-cpu map from.

> For more details:
> http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/concepts_micro_instances.html
>
> I have a few other questions:
>
> 1. Does PCP have any different requirements for running on the cloud.
> 2. Do the metrics need to be interpreted differently when running on cloud.
I don't know enough about the Amazon environment to answer that. Is it
possible for me to login and have a poke around? Or can we download an
instance and run it to investigate?

Regards
-- Mark

> Thanks,
>
> Shirshendu Chakrabarti
>
> On Thu, Nov 27, 2014 at 1:32 PM, Mark Goodwin <mgoodwin@xxxxxxxxxx
> <mailto:mgoodwin@xxxxxxxxxx>> wrote:
>
>     On 11/27/2014 05:17 PM, Shirshendu Chakrabarti wrote:
>
>         Hi PCP team,
>
>         Could anyone please respond the question above :)
>
>
>     The difference is the instance domains and counter aggregation.
>
>     kernel.pernode is aggregated over all CPUs on each numa node.
>     On my laptop with 4 CPU cores on one (fake) numa node, we should
>     expect kernel.all.cpu.user to equal kernel.pernode.cpu.user, and
>     the sum of the per-cpu counters to equal the same, e.g. :
>
>     # pminfo -f kernel.{all,pernode,percpu}.__cpu.user
>
>     kernel.all.cpu.user
>          value 254548080
>
>     kernel.pernode.cpu.user
>          inst [0 or "node0"] value 254548060
>
>     kernel.percpu.cpu.user
>          inst [0 or "cpu0"] value 66377110
>          inst [1 or "cpu1"] value 61187360
>          inst [2 or "cpu2"] value 64051110
>          inst [3 or "cpu3"] value 62932480
>
>     The mapping between CPUs and nodes is in hinv.map.cpu_node,
>     where each CPU (instance) is mapped to a node number, e.g.
>
>     # pminfo -f hinv.map.cpu_node
>
>     hinv.map.cpu_node
>          inst [0 or "cpu0"] value 0
>          inst [1 or "cpu1"] value 0
>          inst [2 or "cpu2"] value 0
>          inst [3 or "cpu3"] value 0
>
>     In your case, kernel.pernode.cpu.user is zero, which isn't correct.
>     What platform and PCP version are you running?
>
>     Regards
>     -- Mark
>
>
>         Thanks,
>
>         Shirshendu Chakrabarti
>
>         On Wed, Nov 26, 2014 at 3:47 PM, Shirshendu Chakrabarti
>         <shirshendu@xxxxxxx <mailto:shirshendu@xxxxxxx>
>         <mailto:shirshendu@xxxxxxx <mailto:shirshendu@xxxxxxx>>> wrote:
>
>              Hi PCP team,
>
>              Could anyone in the team point me to any literature which explains 
> the
>              difference between a per node metric and percpu metric. The
>         one-liner and
>              extended description are absent in pernode metric case and terse in
>         percpu case.
>
>              For example:
>
>              [root@pcp-test-shir3 pmlogger]# pminfo -f kernel.percpu.cpu.user
>
>              kernel.percpu.cpu.user
>                   inst [0 or "cpu0"] value 993240
>              [root@pcp-test-shir3 pmlogger]# pminfo -f kernel.pernode.cpu.user
>
>              kernel.pernode.cpu.user
>                   inst [0 or "node0"] value 0
>              I was under the assumption that, kernel.pernode.* and kernel.all.*
>         metrics
>              are the same. I am clearly missing something important.
>
>              Thanks,
>
>              Shirshendu Chakrabarti
>
>
>
>
>         _________________________________________________
>         pcp mailing list
>         pcp@xxxxxxxxxxx <mailto:pcp@xxxxxxxxxxx>
>         http://oss.sgi.com/mailman/__listinfo/pcp
>         <http://oss.sgi.com/mailman/listinfo/pcp>