Mailing-List: contact mesos-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: mesos-dev@incubator.apache.org
Date: Tue, 5 Jun 2012 03:08:23 +0000 (UTC)
From: "Charles Reiss (JIRA)" <jira@apache.org>
To: mesos-dev@incubator.apache.org
Message-ID: <1199243000.37678.1338865703121.JavaMail.jiratomcat@issues-vm>
In-Reply-To: <234227932.37649.1338864863191.JavaMail.jiratomcat@issues-vm>
Subject: [jira] [Commented] (MESOS-204) enable Frameworks level disk IO
 bandwidth support
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable


    [ https://issues.apache.org/jira/browse/MESOS-204?page=3Dcom.atlassian.=
jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D13289=
108#comment-13289108 ]=20

Charles Reiss commented on MESOS-204:
-------------------------------------

It would probably make more sense to use the cgroups blkio controller (http=
://www.kernel.org/doc/Documentation/cgroups/blkio-controller.txt), assuming=
 you'll have a recent enough kernel. Network prioritization can probably be=
 done similarly. This probably just hasn't been a big priority for people b=
ecause there are nastier issues with isolation, e.g. MESOS-47.

For I/O isolation, one issue is that HDFS (or other DFS) will probably crea=
te priority inversion issues in most real installations. It's not clear how=
 to deal with this problem.
               =20
> enable Frameworks level disk IO bandwidth support
> -------------------------------------------------
>
>                 Key: MESOS-204
>                 URL: https://issues.apache.org/jira/browse/MESOS-204
>             Project: Mesos
>          Issue Type: Brainstorming
>            Reporter: Qinghe Jin
>
> I am considering to add Frameworks level disk IO bandwidth support to mes=
os=EF=BC=8Cwhich behaves like ionice on the single node=EF=BC=8Cbut this ti=
me it=E2=80=98s in multi-nodes=E3=80=82I believe this kind of disk QoS supp=
ort will be more user friendly=E3=80=82My initial idea is to allow user to =
use an ioprio_set like api to set the io priority when they commit their jo=
b. At the same time, here may need some tools to display the disk IO usage =
per node/container on the webui.
> I have read the mesos paper carefully and I found that in the comment=EF=
=BC=8Cit mentioned that in the future it will support disk and network band=
width control=E3=80=82But=EF=BC=8Cin my opinion=EF=BC=8Cdisk IO bandwidth i=
s more feasible for it has more locality=EF=BC=8Cand network bandwidth cont=
rol may be more difficult for the complexity of the network environment=E3=
=80=82But I don't know why is there nobody going to add this feature=E3=80=
=82Is it useless=EF=BC=9FOr not feasible=EF=BC=9FOr just not interested=EF=
=BC=9F
> For this feature may need a lot of work=EF=BC=8CI=E2=80=98d like to liste=
n to your opinions before I start to work on it=E3=80=82Anybody who have an=
y questions ,suggestions or ideas, please tell me and I appreciate it very =
much=EF=BC=8CThanks=EF=BC=81

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrato=
rs: https://issues.apache.org/jira/secure/ContactAdministrators!default.jsp=
a
For more information on JIRA, see: http://www.atlassian.com/software/jira