teuthology consumability

teuthology consumability

[Sage Weil]

Hey all,

I've had a couple different conversations with week with organizations 
that want to ramp up their ceph development efforts, as a big piece of 
that, spin up their own teuthology environments.  There has been a lot 
going on here, and I haven't been on the weekly calls, but I suspect 
things that will help them move forward include:

- documenting a teuthology install-from-scratch.  this would include 
setting up the databases, lock server, workers, etc., and will initially 
be long and ugly but will hopefully shrink over time.

- adding support for euca2ools or some other cloud client that does 
openstack and ec2 for creating/destroying test machines.  hopefully most 
of the integration points are mapped out from the downburst integration 
work.

- ongoing work to collapse the locking and queuing and reporting into a 
unified and/or cohesive service.

I'm probably missing things and/or vastly oversimplifying, but hopefully 
others can chime in an improve this list.  I think the better roadmap we 
have here, the easier it will be for new people to come online and focus 
their efforts where they will make the most difference.

Of course the upcomign CDS is a good place to discuss this, but something 
should be written down first to guide/frame that discussion.

sage

Re: teuthology consumability

[Loic Dachary]

[-- Attachment #1: Type: text/plain, Size: 2118 bytes --]


Hi Sage,

On 07/11/2013 23:21, Sage Weil wrote:
> Hey all,
> 
> I've had a couple different conversations with week with organizations 
> that want to ramp up their ceph development efforts, as a big piece of 
> that, spin up their own teuthology environments.  There has been a lot 
> going on here, and I haven't been on the weekly calls, but I suspect 
> things that will help them move forward include:
> 
> - documenting a teuthology install-from-scratch.  this would include 
> setting up the databases, lock server, workers, etc., and will initially 
> be long and ugly but will hopefully shrink over time.
> 
> - adding support for euca2ools or some other cloud client that does 
> openstack and ec2 for creating/destroying test machines.  hopefully most 
> of the integration points are mapped out from the downburst integration 
> work.
> 

I'm including Hugh Saunders on this thread : we discussed this topic during the OpenStack summit this week in relation to the associated ticket:

http://tracker.ceph.com/issues/6502

> - ongoing work to collapse the locking and queuing and reporting into a 
> unified and/or cohesive service.

It looks like it could be associated with http://wiki.ceph.com/01Planning/02Blueprints/Firefly/Ceph-Brag for performance related tasks. Does that sound sensible ? I'm midly drunk tonight and possibly writing nonsense ;-)

Cheers

> 
> I'm probably missing things and/or vastly oversimplifying, but hopefully 
> others can chime in an improve this list.  I think the better roadmap we 
> have here, the easier it will be for new people to come online and focus 
> their efforts where they will make the most difference.
> 
> Of course the upcomign CDS is a good place to discuss this, but something 
> should be written down first to guide/frame that discussion.
> 
> sage
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

-- 
Loïc Dachary, Artisan Logiciel Libre


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 263 bytes --]

Re: teuthology consumability

[Sage Weil]

On Thu, 7 Nov 2013, Loic Dachary wrote:
> 
> Hi Sage,
> 
> On 07/11/2013 23:21, Sage Weil wrote:
> > Hey all,
> > 
> > I've had a couple different conversations with week with organizations 
> > that want to ramp up their ceph development efforts, as a big piece of 
> > that, spin up their own teuthology environments.  There has been a lot 
> > going on here, and I haven't been on the weekly calls, but I suspect 
> > things that will help them move forward include:
> > 
> > - documenting a teuthology install-from-scratch.  this would include 
> > setting up the databases, lock server, workers, etc., and will initially 
> > be long and ugly but will hopefully shrink over time.
> > 
> > - adding support for euca2ools or some other cloud client that does 
> > openstack and ec2 for creating/destroying test machines.  hopefully most 
> > of the integration points are mapped out from the downburst integration 
> > work.
> > 
> 
> I'm including Hugh Saunders on this thread : we discussed this topic 
> during the OpenStack summit this week in relation to the associated 
> ticket:
> 
> http://tracker.ceph.com/issues/6502

Great.  I would love to see someone sign up for this for the firefly cycle 
(and write up a quick blueprint).  :)
 
> > - ongoing work to collapse the locking and queuing and reporting into a 
> > unified and/or cohesive service.
> 
> It looks like it could be associated with 
> http://wiki.ceph.com/01Planning/02Blueprints/Firefly/Ceph-Brag for 
> performance related tasks. Does that sound sensible ? I'm midly drunk 
> tonight and possibly writing nonsense ;-)

Hmm, I'll go with the latter ;).  I don't think there are tickets open for 
this work, but it's been Zack's main focus these past few weeks.  The last 
glimpse I got of the web interface (just for viewing results, so far) was 
looking pretty slick, too!

sage

> 
> Cheers
> 
> > 
> > I'm probably missing things and/or vastly oversimplifying, but hopefully 
> > others can chime in an improve this list.  I think the better roadmap we 
> > have here, the easier it will be for new people to come online and focus 
> > their efforts where they will make the most difference.
> > 
> > Of course the upcomign CDS is a good place to discuss this, but something 
> > should be written down first to guide/frame that discussion.
> > 
> > sage
> > --
> > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > 
> 
> -- 
> Lo?c Dachary, Artisan Logiciel Libre
> 
> 

Re: teuthology consumability

[Hugh Saunders]

On 8 Nov 2013, at 09:08, "Sage Weil" <sage@inktank.com> wrote:
> 
>> On Thu, 7 Nov 2013, Loic Dachary wrote:
>> 
>> Hi Sage,
>> 
>>> On 07/11/2013 23:21, Sage Weil wrote:
>>> Hey all,
>>> 
>>> I've had a couple different conversations with week with organizations 
>>> that want to ramp up their ceph development efforts, as a big piece of 
>>> that, spin up their own teuthology environments.  There has been a lot 
>>> going on here, and I haven't been on the weekly calls, but I suspect 
>>> things that will help them move forward include:
>>> 
>>> - documenting a teuthology install-from-scratch.  this would include 
>>> setting up the databases, lock server, workers, etc., and will initially 
>>> be long and ugly but will hopefully shrink over time.
>>> 
>>> - adding support for euca2ools or some other cloud client that does 
>>> openstack and ec2 for creating/destroying test machines.  hopefully most 
>>> of the integration points are mapped out from the downburst integration 
>>> work.
>> 
>> I'm including Hugh Saunders on this thread : we discussed this topic 
>> during the OpenStack summit this week in relation to the associated 
>> ticket:
>> 
>> http://tracker.ceph.com/issues/6502
> 
> Great.  I would love to see someone sign up for this for the firefly cycle 
> (and write up a quick blueprint).  :)

Hi, I should have some time to look into teuthology next week. Haven't used it before so will take a while before I can contribute. However notes from setup as a new user should be helpful for docs. 

After that can look at supporting openstack for test instances.

--
Hugh Saunders