Mailing-List: contact issues-help@cloudstack.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cloudstack.apache.org
Date: Tue, 7 Mar 2017 11:58:37 +0000 (UTC)
From: "ASF GitHub Bot (JIRA)" <jira@apache.org>
To: cloudstack-issues@incubator.apache.org
Message-ID: <JIRA.13048844.1488883706000.897.1488887917955@Atlassian.JIRA>
In-Reply-To: <JIRA.13048844.1488883706000@Atlassian.JIRA>
References: <JIRA.13048844.1488883706000@Atlassian.JIRA> <JIRA.13048844.1488883706055@jira-lw-us.apache.org>
Subject: [jira] [Commented] (CLOUDSTACK-9824) Resource count for Primary
 storage is considered twice - while creating and while attaching the disk.
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Tue, 07 Mar 2017 11:58:46 -0000


    [ https://issues.apache.org/jira/browse/CLOUDSTACK-9824?page=3Dcom.atla=
ssian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=
=3D15899326#comment-15899326 ]=20

ASF GitHub Bot commented on CLOUDSTACK-9824:
--------------------------------------------

GitHub user PranaliM opened a pull request:

    https://github.com/apache/cloudstack/pull/1992

    CLOUDSTACK-9824:Resource count for Primary storage is considered twice =
- while creating and while attaching the disk

    The Primary Storage resource counts is considered twice: Once while cre=
ating the volume and next while attaching it. This results in failure while=
 Attaching the volume, if the primary storage resource limit is set.
   =20
    Following are the steps to reproduce the issue:
    1. Create a Domain
    2. Create an Account.
    3. Allocate 11 GB Primary storage limit to the domain.
    4. Deploy a VM sized 8GB (Primary storage resource count is now 8GB )
    5. Create a 2GB custom disk by using the option 'Add Volume' link.(Prim=
ary storage resource count is now 10GB). The Volume state is 'Allocated'
    6. Now try to Attach the data-disk created above to a VM. At this point=
 the resource count limit is again checked, and the attachment fails, becau=
se it is expected to be 12GB, but, which exceeds the limit of 11GB.
    7. This issue is also observed for volumes which are 'Detached'. On det=
aching an attached volume, the state is still seen as 'Ready'. When such a=
=20
    volume is tried to attach again, it will fail as the resource limit wil=
l be checked again.
   =20
    Fix implemented:
    1. When a volume is =E2=80=98Added=E2=80=99, its state is seen as =E2=
=80=98Allocated=E2=80=99 and it utilizes the primary storage resource. When=
 the Volume in =E2=80=98Allocated=E2=80=99 state is attached to a VM, its s=
tate changes to =E2=80=98Ready=E2=80=99.
    2. When the volume is detached from the VM, its state is still seen as =
=E2=80=98Ready=E2=80=99. When the volume is detached, the resources (primar=
y storage) are not released, because the volume is still in primary storage=
.=20
    3. The fix is to skip checking the resource limit when the volume state=
 is either 'Allocated' or 'Ready'. Because the volume state being 'Allocate=
d' or 'Ready' implies that it is already on the Primary storage and hence t=
he limit should not be checked again.
   =20
    Test Results:
    1. Set a limit of 42 GB primary storage to a new domain=20
    2. Deployed a 30GB VM in it.=20
    3. Created a 10 GB disk and could attach and detach it multiple times w=
ithout any error of resource limit.

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/Accelerite/cloudstack CS-9824

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/cloudstack/pull/1992.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #1992
   =20
----
commit 616c05e2921af6b713f8a4bce2d2421b03462a22
Author: Pranali Mande <pranali_mande@accelerite.com>
Date:   2017-03-07T11:53:02Z

    CLOUDSTACK-9824:Resource count for Primary storage is considered twice =
- while creating and while attaching the disk

----


> Resource count for Primary storage is considered twice - while creating a=
nd while attaching the disk.
> -------------------------------------------------------------------------=
----------------------------
>
>                 Key: CLOUDSTACK-9824
>                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-982=
4
>             Project: CloudStack
>          Issue Type: Bug
>      Security Level: Public(Anyone can view this level - this is the defa=
ult.)=20
>            Reporter: Pranali Mande
>
> The Primary Storage resource counts is considered twice: Once while creat=
ing the volume and next while attaching it. This results in failure while A=
ttaching the volume, if the primary storage resource limit is set.
> Following are the steps to reproduce the issue:
> 1. Create a Domain
> 2. Create an Account.
> 3. Allocate 11 GB Primary storage limit to the domain.
> 4. Deploy a VM sized 8GB (Primary storage resource count is now 8GB )
> 5. Create a 2GB custom disk by using the option 'Add Volume' link.(Primar=
y storage resource count is now 10GB). The Volume state is 'Allocated'
> 6. Now try to Attach the data-disk created above to a VM. At this point t=
he resource count limit is again checked, and the attachment fails, because=
 it is expected to be 12GB, but, which exceeds the limit of 11GB.
> 7. This issue is also observed for volumes which are 'Detached'. On detac=
hing an attached volume, the state is still seen as 'Ready'. When such a=20
> volume is tried to attach again, it will fail as the resource limit will =
be checked again.
> Fix implemented:
> 1. When a volume is =E2=80=98Added=E2=80=99, its state is seen as =E2=80=
=98Allocated=E2=80=99 and it utilizes the primary storage resource. When th=
e Volume in =E2=80=98Allocated=E2=80=99 state is attached to a VM, its stat=
e changes to =E2=80=98Ready=E2=80=99.
> 2. When the volume is detached from the VM, its state is still seen as =
=E2=80=98Ready=E2=80=99. When the volume is detached, the resources (primar=
y storage) are not released, because the volume is still in primary storage=
.=20
> 3. The fix is to skip checking the resource limit when the volume state i=
s either 'Allocated' or 'Ready'. Because the volume state being 'Allocated'=
 or 'Ready' implies that it is already on the Primary storage and hence the=
 limit should not be checked again.
> Test Results:
> 1. Set a limit of 42 GB primary storage to a new domain=20
> 2. Deployed a 30GB VM in it.=20
> 3. Created a 10 GB disk and could attach and detach it multiple times wit=
hout any error of resource limit.


--
This message was sent by Atlassian JIRA
(v6.3.15#6346)