thanks for your answer, i will be test and report to you,
On 09/03/2014 04:09 PM, Javier Fontan wrote:
We have not tested that configuration but if you only use persistent images it may work. Make sure you don't try to do cold migrations as the checkpoint is saved to the system datastore and probably will get some time to be synchronized with the new node.


On Tue, Aug 26, 2014 at 9:04 AM, Rhesa Mahendra <rh...@lintasmediadanawa.com <mailto:rh...@lintasmediadanawa.com>> wrote:

    Datastore 0 and datastore 2 will be synchronized frontend + nodes
    (we just use persistent image) so i think datastore 0 just
    synchronized symbolic link.
    If we have problem maybe existing storage box (datastore 1)
    broken, all vm used new storage box ( datastore 100) will not
    impact with datastore 1 because datastore 0 shared and
    synchronized across frontend +nodes. it possible?

    Rhesa


    On 08/26/2014 01:52 PM, Jaime Melis wrote:
    Rhesa,

    the problem is datastore 0, it's the system datastore. If you
    move it to the local nodes, it will not be shared and
    synchronized across the nodes, therefore you will lose many
    benefits like live-migration and instant persintant image VM
    deployment.

    Why is it that you want to move datastore 0 to the frontend + nodes?


    On Tue, Aug 26, 2014 at 8:44 AM, Rhesa Mahendra
    <rh...@lintasmediadanawa.com
    <mailto:rh...@lintasmediadanawa.com>> wrote:

        Jaime,

        Thanks for your reply, now datastore 0 and datastore 2 are in
        existing storage box, mounted in frontend + nodes, we want to
        move datastore 0 and datastore 2 from existing storage box to
        frontend + nodes, could it be done?

        So in existing storage box just have datastore 1 and in new
        storage box just have datastore 100.

        Rhesa.


        On 08/26/2014 01:34 PM, Jaime Melis wrote:
        Hi,

        I don't entirely understand the difference between both
        options, could you explain in more detail what you mean by
        distributing the datastore to all the nodes?

        Assuming your configuration is as follows: NFS Server
        exports datastores 0 and 1 -> Mounted in Frontend + Nodes

        I recommend the following option:
        - Create a new Image datastore (it will have ID >= 100, in
        this example we will assume it has ID: 100)
        - Create a new exports directive in your new storage box,
        for example /storage/datastores/100
        - Mount that mountpoint in all the nodes including the
        frontend in the following path: /var/lib/one/datastores/100

        Now you can register more images in the new datastore (ID:
        100), and since they are all persistent they will not take
        space in the system datastore (ID: 0)

        Best regards,
        Jaime



        On Tue, Aug 26, 2014 at 6:05 AM, Rhesa Mahendra
        <rh...@lintasmediadanawa.com
        <mailto:rh...@lintasmediadanawa.com>> wrote:

            Guys, please help.



            -------- Original Message --------
            Subject:    Add New Storage Box
            Date:       Mon, 25 Aug 2014 16:21:13 +0700
            From:       Rhesa Mahendra <rh...@lintasmediadanawa.com>
            <mailto:rh...@lintasmediadanawa.com>
            To:         users@lists.opennebula.org
            <mailto:users@lists.opennebula.org>
            <users@lists.opennebula.org>
            <mailto:users@lists.opennebula.org>, Javier Fontan
            <jfon...@opennebula.org> <mailto:jfon...@opennebula.org>



            Guys,

            We have plan to add a new storage box, Our existing storage :
            /var/lib/one/datastore/0 = system
            /var/lib/one/datastore/1 = image
            /var/lib/one/datastore/2 = file
            All type are from one storage box.

            We did internal discuss and we get result with two options:

            1. We add a new storage box as an image type and define it as
            /var/lib/one/datastore/3 = image2.

            Or

            2. We will distribute /var/lib/one/datastore/0 = system and
            /var/lib/one/datastore/2 = file to all nodes (all node are synced), 
then
            we add the new storage box as /var/lib/one/datastore/3 = image2
            So with this second option, the nodes only mount two path from the
            storage box (/var/lib/one/datastore/1 = image [from existing storage
            box]  and /var/lib/one/datastore/3 = image2 [from new storage 
box]). All
            the images are persistent
            Please give us some advise with pros and cons within this two 
options,
            thanks.

            Rhesa




            _______________________________________________
            Users mailing list
            Users@lists.opennebula.org
            <mailto:Users@lists.opennebula.org>
            http://lists.opennebula.org/listinfo.cgi/users-opennebula.org




-- Jaime Melis
        Project Engineer
        OpenNebula - Flexible Enterprise Cloud Made Simple
        www.OpenNebula.org <http://www.OpenNebula.org> |
        jme...@opennebula.org <mailto:jme...@opennebula.org>


        _______________________________________________
        Users mailing list
        Users@lists.opennebula.org <mailto:Users@lists.opennebula.org>
        http://lists.opennebula.org/listinfo.cgi/users-opennebula.org




-- Jaime Melis
    Project Engineer
    OpenNebula - Flexible Enterprise Cloud Made Simple
    www.OpenNebula.org <http://www.OpenNebula.org> |
    jme...@opennebula.org <mailto:jme...@opennebula.org>


    _______________________________________________
    Users mailing list
    Users@lists.opennebula.org <mailto:Users@lists.opennebula.org>
    http://lists.opennebula.org/listinfo.cgi/users-opennebula.org




--
Javier Fontán Muiños
Developer
OpenNebula - Flexible Enterprise Cloud Made Simple
www.OpenNebula.org <http://www.OpenNebula.org> | @OpenNebula | github.com/jfontan <http://github.com/jfontan>

_______________________________________________
Users mailing list
Users@lists.opennebula.org
http://lists.opennebula.org/listinfo.cgi/users-opennebula.org

Reply via email to