[LSF/MM TOPIC] Springfield: smart and automated storage

[LSF/MM TOPIC] Springfield: smart and automated storage

[Jan Tulak]

Springfield is a collection of projects unifying multiple levels of
the storage stack and providing a general API for automation, health
and status monitoring, as well as sane and easy configuration across
multiple levels of the storage stack. It is a scalable solution,
working from a single node to large deployments with a mix of base
metal, containers and VMs. For the first time, it was presented on
Vault 2017 in a birds of a feather session.

Springfield builds upon and enhances the existing work of those projects:
* udisks
* libblockdev
* blivet
* libstoragemgmt

It is a coordinated effort to overcome many of the shortcomings of the
current situation and provide a unified approach to storage
management, so there is nothing like a binary or a library with the
name “Springfield.”

An example of things we are trying to tackle is a useful storage
reporting, like notifications about filesystem being full, especially
with thin provisioning and multiple levels of the storage stack, where
the user-visible status is not a true representation of what is
happening underneath.

We want to see how the community sees this effort. What shortcomings
of the current situation in storage layering can Springfield address
(is there something we don’t see, but a project like this could help)?
How would you use it, or what would make it more useful to you?

Best regards,
Jan Tulak

Re: [LSF/MM TOPIC] Springfield: smart and automated storage

[James Bottomley]

On Sat, 2018-01-27 at 09:37 +0100, Jan Tulak wrote:
> Springfield is a collection of projects unifying multiple levels of
> the storage stack and providing a general API for automation, health
> and status monitoring, as well as sane and easy configuration across
> multiple levels of the storage stack. It is a scalable solution,
> working from a single node to large deployments with a mix of base
> metal, containers and VMs. For the first time, it was presented on
> Vault 2017 in a birds of a feather session.
> 
> Springfield builds upon and enhances the existing work of those
> projects:
> * udisks
> * libblockdev
> * blivet
> * libstoragemgmt
> 
> It is a coordinated effort to overcome many of the shortcomings of
> the current situation and provide a unified approach to storage
> management, so there is nothing like a binary or a library with the
> name “Springfield.”
> 
> An example of things we are trying to tackle is a useful storage
> reporting, like notifications about filesystem being full, especially
> with thin provisioning and multiple levels of the storage stack,
> where
> the user-visible status is not a true representation of what is
> happening underneath.
> 
> We want to see how the community sees this effort. What shortcomings
> of the current situation in storage layering can Springfield address
> (is there something we don’t see, but a project like this could
> help)? How would you use it, or what would make it more useful to
> you?

Do you have a link to the project page and the source tree(s)?

James

Re: [LSF/MM TOPIC] Springfield: smart and automated storage

[Jan Tulak]

On Sat, Jan 27, 2018 at 5:32 PM, James Bottomley
<James.Bottomley@hansenpartnership.com> wrote:
> On Sat, 2018-01-27 at 09:37 +0100, Jan Tulak wrote:
>> Springfield is a collection of projects unifying multiple levels of
>> the storage stack and providing a general API for automation, health
>> and status monitoring, as well as sane and easy configuration across
>> multiple levels of the storage stack. It is a scalable solution,
>> working from a single node to large deployments with a mix of base
>> metal, containers and VMs. For the first time, it was presented on
>> Vault 2017 in a birds of a feather session.
>>
>> Springfield builds upon and enhances the existing work of those
>> projects:
>> * udisks
>> * libblockdev
>> * blivet
>> * libstoragemgmt
>>
>> It is a coordinated effort to overcome many of the shortcomings of
>> the current situation and provide a unified approach to storage
>> management, so there is nothing like a binary or a library with the
>> name “Springfield.”
>>
>> An example of things we are trying to tackle is a useful storage
>> reporting, like notifications about filesystem being full, especially
>> with thin provisioning and multiple levels of the storage stack,
>> where
>> the user-visible status is not a true representation of what is
>> happening underneath.
>>
>> We want to see how the community sees this effort. What shortcomings
>> of the current situation in storage layering can Springfield address
>> (is there something we don’t see, but a project like this could
>> help)? How would you use it, or what would make it more useful to
>> you?
>
> Do you have a link to the project page and the source tree(s)?
>

Sure, for Springfield the website is: https://springfield-project.github.io/

Each of the listed projects has it's own repositories:
https://github.com/storaged-project/udisks
https://github.com/libstorage/libstoragemgmt
https://github.com/storaged-project/libblockdev
https://github.com/storaged-project/blivet

Cheers,
Jan