filesystem type storage backend

[Marco Ceppi]

Hello,

We had a discussion earlier today about how the filesystem backend works
for the storage bits. Given the following example:

storage:
  hdfs-data:
type: filesystem
location: /srv/hdfs

Does this require the operator to add storage at deploy time? Does this
mean our charm should block until storage is added? We really want to be
able to support the /option/ or adding storage by the operator but we're
not sure how this functions. If we use that directory before the operator
attaches storage will it be mounted over or migrated?

Thanks,
Marco Ceppi
-- 
Juju-dev mailing list
Juju-dev@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/juju-dev

Re: filesystem type storage backend

[Marco Ceppi]

This sounds very promising! Will play around with it a bit more.

On Fri, Feb 5, 2016, 7:16 PM Mark Shuttleworth  wrote:

> On 05/02/16 17:47, Marco Ceppi wrote:
> > Does this require the operator to add storage at deploy time? Does this
> > mean our charm should block until storage is added? We really want to be
> > able to support the /option/ or adding storage by the operator but we're
> > not sure how this functions. If we use that directory before the operator
> > attaches storage will it be mounted over or migrated?
>
> AIUI:
>
>  * you can leave the storage unspecified and it will just write to the
> directory
>  * you cannot change the storage pool after deployment
>
> Mark
>
-- 
Juju-dev mailing list
Juju-dev@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/juju-dev