Mailing-List: contact dev-help@cloudstack.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cloudstack.apache.org
Received-SPF: pass (nike.apache.org: domain of shadowsor@gmail.com designates
 209.85.220.181 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <BLU184-W800128E9F1CFEB1B0F769AB4300@phx.gbl>
References: <BLU184-W48D0E770BACFEF8E8BCBAFB4300@phx.gbl>
	<D25E116F470C5E4A869A2A9E428836322455AB10@SINPEX01CL01.citrite.net>
	<BLU184-W55D07DFD28FAA050768342B4300@phx.gbl>
	<CALFpzo5=9eTaUKLwP=wuOG8ABNJBQnpoh-Q06XX9SHGqTLO1SQ@mail.gmail.com>
	<BLU184-W800128E9F1CFEB1B0F769AB4300@phx.gbl>
Date: Thu, 29 Jan 2015 00:24:35 -0700
Message-ID: 
 <CALFpzo7X9a=WTwJtiCQsGzEOEfgtMfFdgFL=uteXjCsS2ZS7Nw@mail.gmail.com>
Subject: Re: Orphaned Data Disks
From: Marcus <shadowsor@gmail.com>
To: "dev@cloudstack.apache.org" <dev@cloudstack.apache.org>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

The trick is to not think of them as orphaned. The data disk is its
own entity, it stands alone from the VM, and really the ability to
create a data disk along with vm deployment is just a convenience,
data disk lifecycle is to be created standalone, attached, detached,
and deleted standalone.

It's a difference between standard VM management and 'cloud' style
'instance' management. In AWS your VM is essentially recreated every
time you reboot, and the only way to have persistent data is with a
datadisk (or object store/separate db) that gets attached to your new
VM each time.  Luckily, CloudStack is a good bridge for that gap,
being friendly to the more 'old school' VM management where instances
are pets with things like persistent root disks, but built with cloud
workloads in mind where you don't care about instances but still want
persistent datastores that outlive the instance.

That said, I do agree that it wouldn't be a big deal to add a
parameter to the destroyVirtualMachine api call that would cause it to
loop through all attached disks and remove them. You can open a
feature request at https://issues.apache.org

On Thu, Jan 29, 2015 at 12:01 AM, Michael Phillips
<mphilli7823@hotmail.com> wrote:
> To me that's kind of strange to NOT assume to delete the disks that are a=
ttached to an instance at time of deletion.  I think in most real world env=
ironments a data disk will belong to one machine only. Of course the except=
ion is clustering and that is probably outside of this scope. Even if I was=
 going to do some kind of OS upgrade and wanted to reuse the data disk on a=
nother instance I would probably detach, then reattach to the new instance.=
 Just seems like it can get messy quick, if a lot of users delete their ins=
tances and leave all these orphaned data disks behind.
> It would be awesome to have a selection box to have cloudstack delete ALL=
 attached disks when an instance is destroyed. Just my 2 cents...
>
>> Date: Wed, 28 Jan 2015 22:56:43 -0700
>> Subject: Re: Orphaned Data Disks
>> From: shadowsor@gmail.com
>> To: dev@cloudstack.apache.org
>>
>> Data disks are their own entity. You can detach them and attach them
>> to other VMs. CloudStack doesn't assume that you want all the disks to
>> die when you destroy a VM simply because they happen to be attached to
>> that vm at the moment.
>>
>> On Wed, Jan 28, 2015 at 10:47 PM, Michael Phillips
>> <mphilli7823@hotmail.com> wrote:
>> > What was the logic behind leaving the disks orphaned?
>> >
>> >> From: sanjeev.neelarapu@citrix.com
>> >> To: dev@cloudstack.apache.org
>> >> Subject: RE: Orphaned Data Disks
>> >> Date: Thu, 29 Jan 2015 05:30:49 +0000
>> >>
>> >> That is expected behavior. Right now there is no option to change it.
>> >>
>> >> Sanjeev,
>> >> CloudPlatform Engineering,
>> >> Citrix Systems, Inc.
>> >>
>> >>
>> >> -----Original Message-----
>> >> From: Michael Phillips [mailto:mphilli7823@hotmail.com]
>> >> Sent: Thursday, January 29, 2015 10:43 AM
>> >> To: dev@cloudstack.apache.org
>> >> Subject: Orphaned Data Disks
>> >>
>> >> Has anyone noticed that after destroying and expunging an instance th=
at has a data disk attached, cloudstack leaves the datadisk orphaned? If so=
 is this expected behavior, and if so is there an option to change it?
>> >
>