Mailing-List: contact issues-help@cloudstack.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cloudstack.apache.org
Date: Thu, 2 Jul 2015 09:41:07 +0000 (UTC)
From: "Likitha Shetty (JIRA)" <jira@apache.org>
To: cloudstack-issues@incubator.apache.org
Message-ID: <JIRA.12842267.1435830065000.83533.1435830067213@Atlassian.JIRA>
In-Reply-To: <JIRA.12842267.1435830065000@Atlassian.JIRA>
References: <JIRA.12842267.1435830065000@Atlassian.JIRA>
 <JIRA.12842267.1435830065626@arcas>
Subject: [jira] [Created] (CLOUDSTACK-8612) [VMware] Make vCenter session
 timeout configurable for volume snapshot
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

Likitha Shetty created CLOUDSTACK-8612:
------------------------------------------

             Summary: [VMware] Make vCenter session timeout configurable fo=
r volume snapshot
                 Key: CLOUDSTACK-8612
                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-8612
             Project: CloudStack
          Issue Type: Bug
      Security Level: Public (Anyone can view this level - this is the defa=
ult.)
            Reporter: Likitha Shetty
            Assignee: Likitha Shetty
             Fix For: 4.6.0


+Issue+
Taking a volume snapshot involves the below steps in case of VMware -=C2=A0
1. Create a VM snapshot for the VM that the volume is attached to.
2. Create a VM (clone) from the snapshot disk chain of the volume to be sna=
pshot'ed.
3. Export the VM created in step 2 to secondary storage.
4. Destroy the cloned VM.
5. Delete the VM snapshot taken in step 1 (This step involves VM consolidat=
ion in vCenter).

If any of the above steps (except VM export) takes more than 20 minutes to =
complete, snapshot backup operation will fail and the snapshot will be move=
d to 'Error' state. 20 minutes because the session between CS and vCenter t=
imes out in 20 minutes. =20

In case of VM's with large volumes (most production environments) because o=
f the size of the volumes and the fact that vCenter takes too long to conso=
lidate VMs this large (step 5), snapshot operation on the volume attached t=
o this VM takes too long to complete even going beyond a day.=20
Snapshot consolidation taking too long is a known issue with vCenter - [htt=
p://kb.vmware.com/selfservice/search.do?cmd=3DdisplayKC&docType=3Dkc&docTyp=
eID=3DDT_KB_1_1&externalId=3D1023657]

+Solution+
To work around this, CS=E2=80=99s session to vCenter during snapshot operat=
ion needs to be set to a value higher than the time taken to consolidate fo=
r the operation to successfully complete. But setting the existing global c=
onfiguration to this high a value could affect all operations in the cloud =
and hence is not considered an ideal workaround.

Hence we need to split the existing vCenter session timeout configuration i=
nto two configurations,
1. To tune timeout value for CCP=E2=80=99s session vCenter during snapshot =
backup operation - _vmware.snapshot.backup.session.timeout_ (new; default: =
20 minutes)
2. For all other operations - _vmware.vcenter.session.timeout_ (existing; d=
efault: 20 minutes)

This way setting a high timeout value for snapshot backup operation will no=
t be applicable to vCenter connections made for any other operations.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)