Calamari 1.4

Calamari 1.4

[Gregory Meno]

Coming soon to a jewel near you.

What is it:

Calamari is a REST API written in python. It seeks to give API
endpoints like /cluster,
/pool, and /osd and unsurprisingly allow HTTP requests to inspect and
modify the resources a Ceph cluster provides. It simplifies access to
the control-plane by making it available across the network.

What is changing from the calamari I knew:
Calamari is going on a diet. We're shedding dependencies on saltstack.
The new version will be able to be run along-side Ceph Monitor
processes on the monitor nodes.  Where before calamari sought to place
agents on all nodes on the clusters it managed and monitored;  we now
only care about the cluster we're setup on

What is staying the same:
The API -- http://calamari.readthedocs.org/en/latest/calamari_rest/index.html
This is really the part I'm most excited about. If you've used
calamari in the past I am committed to delivering compatibility in
this release


How can I try it out:
* note this is still development code -- don't put it on production clusters
* http://calamari.readthedocs.org/en/latest/development/dev_env.html#using-vagrant
* release candidate packages will be available soon


How can I help:
* Contribute doc improvements
* File bugs in https://github.com/ceph/calamari/issues
* Fix bugs :)
* help with packaging
* contribute to the future direction by discussing this on the mailing
list -- ceph-calamari@lists.ceph.com

regards,
Gregory

Re: [ceph-calamari] Calamari 1.4

[Daniel Rolfe]

Hi Gregory , 

Have you considered doing away with the graphite server in favor for another metrics db like influxdb 

Just doesn't make sense to install all the graphite dependencies just to pull metric data

Plus do we really need two web interfaces i.e calamari web and graphite web

Just my thoughts

Regards Daniel

> On 20 Feb 2016, at 12:14 PM, Gregory Meno <gmeno@redhat.com> wrote:
> 
> Coming soon to a jewel near you.
> 
> What is it:
> 
> Calamari is a REST API written in python. It seeks to give API
> endpoints like /cluster,
> /pool, and /osd and unsurprisingly allow HTTP requests to inspect and
> modify the resources a Ceph cluster provides. It simplifies access to
> the control-plane by making it available across the network.
> 
> What is changing from the calamari I knew:
> Calamari is going on a diet. We're shedding dependencies on saltstack.
> The new version will be able to be run along-side Ceph Monitor
> processes on the monitor nodes.  Where before calamari sought to place
> agents on all nodes on the clusters it managed and monitored;  we now
> only care about the cluster we're setup on
> 
> What is staying the same:
> The API -- http://calamari.readthedocs.org/en/latest/calamari_rest/index.html
> This is really the part I'm most excited about. If you've used
> calamari in the past I am committed to delivering compatibility in
> this release
> 
> 
> How can I try it out:
> * note this is still development code -- don't put it on production clusters
> * http://calamari.readthedocs.org/en/latest/development/dev_env.html#using-vagrant
> * release candidate packages will be available soon
> 
> 
> How can I help:
> * Contribute doc improvements
> * File bugs in https://github.com/ceph/calamari/issues
> * Fix bugs :)
> * help with packaging
> * contribute to the future direction by discussing this on the mailing
> list -- ceph-calamari@lists.ceph.com
> 
> regards,
> Gregory
> _______________________________________________
> ceph-calamari mailing list
> ceph-calamari@lists.ceph.com
> http://lists.ceph.com/listinfo.cgi/ceph-calamari-ceph.com

Re: [ceph-calamari] Calamari 1.4

[Tim Serong]

On 02/20/2016 02:14 AM, Gregory Meno wrote:
> Coming soon to a jewel near you.
> 
> What is it:
> 
> Calamari is a REST API written in python. It seeks to give API
> endpoints like /cluster,
> /pool, and /osd and unsurprisingly allow HTTP requests to inspect and
> modify the resources a Ceph cluster provides. It simplifies access to
> the control-plane by making it available across the network.
> 
> [...]
> 
> How can I help:
> * Contribute doc improvements
> * File bugs in https://github.com/ceph/calamari/issues
> * Fix bugs :)
> * help with packaging

I can help with that (as of a couple of days ago I've got a very rough
RPM build for SLES and openSUSE).  You'll want patches against the
calamari_on_monitor branch, right?  Or is there going to be a 1.4 branch?

Regards,

Tim
-- 
Tim Serong
Senior Clustering Engineer
SUSE
tserong@suse.com

Re: [ceph-calamari] Calamari 1.4

[Gregory Meno]

On Fri, Feb 19, 2016 at 5:32 PM, Daniel Rolfe <daniel.rolfe.au@gmail.com> wrote:
> Hi Gregory ,
>
> Have you considered doing away with the graphite server in favor for another metrics db like influxdb
>
> Just doesn't make sense to install all the graphite dependencies just to pull metric data
>
> Plus do we really need two web interfaces i.e calamari web and graphite web
>
> Just my thoughts
>
> Regards Daniel
>
>> On 20 Feb 2016, at 12:14 PM, Gregory Meno <gmeno@redhat.com> wrote:
>>
>> Coming soon to a jewel near you.
>>
>> What is it:
>>
>> Calamari is a REST API written in python. It seeks to give API
>> endpoints like /cluster,
>> /pool, and /osd and unsurprisingly allow HTTP requests to inspect and
>> modify the resources a Ceph cluster provides. It simplifies access to
>> the control-plane by making it available across the network.
>>
>> What is changing from the calamari I knew:
>> Calamari is going on a diet. We're shedding dependencies on saltstack.
>> The new version will be able to be run along-side Ceph Monitor
>> processes on the monitor nodes.  Where before calamari sought to place
>> agents on all nodes on the clusters it managed and monitored;  we now
>> only care about the cluster we're setup on
>>
>> What is staying the same:
>> The API -- http://calamari.readthedocs.org/en/latest/calamari_rest/index.html
>> This is really the part I'm most excited about. If you've used
>> calamari in the past I am committed to delivering compatibility in
>> this release
>>
>>
>> How can I try it out:
>> * note this is still development code -- don't put it on production clusters
>> * http://calamari.readthedocs.org/en/latest/development/dev_env.html#using-vagrant
>> * release candidate packages will be available soon
>>
>>
>> How can I help:
>> * Contribute doc improvements
>> * File bugs in https://github.com/ceph/calamari/issues
>> * Fix bugs :)
>> * help with packaging
>> * contribute to the future direction by discussing this on the mailing
>> list -- ceph-calamari@lists.ceph.com
>>
>> regards,
>> Gregory
>> _______________________________________________
>> ceph-calamari mailing list
>> ceph-calamari@lists.ceph.com
>> http://lists.ceph.com/listinfo.cgi/ceph-calamari-ceph.com

Daniel,

I'm not sure I understand what you are asking. We could certainly try
to strip it down now that I'm not really expecting graphs output for a
GUI.

Graphite is really just a front-end to the data living in carbon and
reported by diamond.

The sticking point is that I believe that some are using the graphite
API that is wrapped by calamari. Phillip would you please say more
about this?

Influxdb seems promising AND I have a concern that it seems to not be
packages in any of the distros yet.

https://admin.fedoraproject.org/pkgdb/packages/?motif=influxdb*

http://packages.ubuntu.com/search?keywords=influxdb -- sure it's in
xenial but I'm not there yet

https://software.opensuse.org/package/influxdb


What do you think?

thank you,
Gregory