pcp
[Top] [All Lists]

Re: PCP 3.10.5

To: Nathan Scott <nathans@xxxxxxxxxx>
Subject: Re: PCP 3.10.5
From: henry c <hc3249@xxxxxxxxx>
Date: Thu, 30 Apr 2015 16:45:22 -0700
Cc: pcp developers <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=eMqXtcYg0pvfJtXeZ/o2RYh61CvmwS0cQ9/YAIHcIzY=; b=PFiZwh9fY3BuvIzSOKemfTMzqDNskDr9h+GgY8h8Dxnji7toy8yK8MLiQ76KD83yv8 16iCfso1Y27S3ozz6n9qDoyryGdK63iIYoB4NSHuBTW+UF1T2X2KtLpwnhvuV0RSoZOF s+WyFakxP7UwEuj+ZXU5OWTwEAKiQKBnKP19Bc58pLdwg0F69WVsVrM93rN57f6buods vDsXkqHKbEBxRNkFIocWrRNwgLGDcgbfCpqED9jbSY6RRNVkoqc9mFqRvnuhEY2GbrA7 TIqz4atl0SBxhymg0mWrZlKNhny9f1msbrUtXpLwSJvLJYR1gexv43fbdaggpbx5zErb hZBg==
In-reply-to: <1853303344.10742742.1430435378053.JavaMail.zimbra@xxxxxxxxxx>
References: <CABMF9cuiUXe3Sss+ZUZn4oh1Stsp43u1G1ib_DQhZ58v1FimdA@xxxxxxxxxxxxxx> <513687800.9959776.1430348643936.JavaMail.zimbra@xxxxxxxxxx> <CABMF9cuiVbz83dHNMb2kyBUG=SBc8=hMtnH3=c52YwP=ONasaw@xxxxxxxxxxxxxx> <1853303344.10742742.1430435378053.JavaMail.zimbra@xxxxxxxxxx>
Nathan,

Thanks for the response.

For me at least, the numbers are not matching.


CONTAINER IDÂ Â Â Â IMAGE Â Â Â Â Â Â Â COMMANDÂ Â Â Â Â Â Â Â CREATED Â Â Â Â Â Â STATUSÂ Â Â Â Â Â Â PORTS Â Â Â Â Â Â Â NAMES

5a86bb5685bc    fbe679c040d6    "/bin/sh -c '/usr/lo  2 weeks ago     Up 2 weeks               serene_pike


host:/sys/fs/cgroup/memory/docker/5a86bb5685bcfc87159f2578079703efc172c662ac9702757f1829b53d15d8f6#Âcat memory.max_usage_in_bytes

1822720


and calling the api returns this, with all instance having the same value


    {

      "pmid": 12629073,

      "name": "cgroup.memory.limit",

      "instances": [

        {

          "instance": 0,

          "value": 18446744073709552000

        },

        {


Thanks,

-Henry


On Thu, Apr 30, 2015 at 4:09 PM, Nathan Scott <nathans@xxxxxxxxxx> wrote:
Hi there Henry,

(CC'ing the PCP list in case others come across this phenomenon too)

----- Original Message -----
> Hey Nathan,
>
> Thanks for the quick response!
>
> I was able to install the new code, and get data from those 3 new metrics.
>
> However the cgroup.memory.limit looks weird to me, i dont think this is in
> bit or bytes since it would be too large.

Thanks for trying it out! These values are sourced from the kernel via the
cgroup pseudo-filesystem, and we pass back (unmodified) the contents of the
statistics files for each cgroup.

You should be able to find these files and check what limits/usage the kernel
is applying. For cgroups that are used in Docker containers, you should see
things like:

$ cat /sys/fs/cgroup/memory/system.slice/docker-c4f1a5c72ee11e819c1560b6a3be752d3dd6486f84df8322647ad96a184f1f4d.scope/memory.max_usage_in_bytes
14336000
$ cat /sys/fs/cgroup/memory/system.slice/docker-c4f1a5c72ee11e819c1560b6a3be752d3dd6486f84df8322647ad96a184f1f4d.scope/memory.usage_in_bytes
528384

$ docker ps | grep c4f1a5c72ee
c4f1a5c72ee1    fedora:20     Â"/bin/bash"    Â18 hours ago    Up 18 hours              Âjovial_galileo

$ pminfo --fetch --container=jovial_galileo cgroup.memory.usage

cgroup.memory.usage
  inst [0 or "/system.slice/docker-c4f1a5c72ee11e819c1560b6a3be752d3dd6486f84df8322647ad96a184f1f4d.scope"] value 528384


cheers.

--
Nathan

<Prev in Thread] Current Thread [Next in Thread>