incubator-cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Edison Su <Edison...@citrix.com>
Subject RE: Local storage support for data volumes
Date Tue, 14 Aug 2012 00:53:08 GMT


> -----Original Message-----
> From: Matthew Patton [mailto:mpatton@inforelay.com]
> Sent: Monday, August 13, 2012 6:47 AM
> To: cloudstack-dev@incubator.apache.org
> Subject: RE: Local storage support for data volumes
> 
> Unfortunately this caught my eye.
> " n disabled existing local storages are not removed but any new local
> storage is not added."
> 
> This qualifies as interference. Cloud software must always reflect
> reality. This whole distinction between root or data being allowed as
> local was none of CS' business in the first place.
> 
> Of course CS should throw an exception if you try to live migrate a
> local storage based instance. it can even refuse to do a cold migrate
> until such time as the needed code is written.
> 
> But it is uncalled for CS to disallow otherwise legal operations or to
> unilaterally hide key elements because some programmer without
> sufficient real world experience thinks the world should fit their
> preconceived notions.

I agree, CS should not hide local storage. We can enable it or disable it, but should never
hide it.
In current implementation, if use.local.storage is false, then the local storage will not
be added into cloudstack database at all, thus UI will not show it, admin will not see it.

> 
> Furthermore all three hypervisor flavors support movement of local VMs,
> not just Xen.
> 
> First rule of UI design, do not lie to the user by hiding things. First
> rule of software development, programmers are generally horrible at
> engineering.

Mime
View raw message