Re: [openstack-dev] [CINDER] [PTL Candidates] Questions

[Walter A. Boring IV]

>
> 1. Do you actually have the time to spend to be PTL
>
> I don't think many people realize the time commitment. Between being
> on top of reviews and having a pretty consistent view of what's going
> on and in process; to meetings, questions on IRC, program management
> type stuff etc.  Do you feel you'll have the ability for PTL to be
> your FULL Time job?  Don't forget you're working with folks in a
> community that spans multiple time zones.
The short answer to this is yes.   Prior to even putting up my candidacy
I spoke with my management and informed them of what would be involved
with being PTL for Cinder, and that meant it was an upstream job.  I've
been working on Cinder for 3 years now and have seen the amount of time
that you and Mike have spent on the project, and it's significant to say
the least.   The wiki has a good guide for PTL candidates here:
https://wiki.openstack.org/wiki/PTL_Guide.   It's a decent start and
more of a "PTL for dummies" guide and is by no means everything a PTL is
and has to do.  Being a PTL means more than just attending meetings,
doing reviews, and communication.  It means being the lead evangelist
and ambassador for Cinder.   As PTL of a project, it's also important
not to forget about the future of the community and encourage new
members to contribute code to Cinder core itself, to help make Cinder a
better project.  For example, the recent additions by Kendall Nelson to
work on the cinder.conf.sample file
(https://review.openstack.org/#/c/219700).  The patch itself might have
more follow up work, as noted in the review, but she was very responsive
and was on top of the code to try and get it to land.  Sean, John and
myself all helped with reviews on that patch and worked together as a
team to help Kendall with her efforts.  We need more new contributors
like her.  The more inclusive and encouraging of new members in the
community the better.   I remember starting out working on Cinder back
in the Grizzly time frame and I also remember John, as the PTL, being
very helpful and encouraging of my efforts to learn how to write a
driver and how to contribute in general.  It was a very welcoming
experience at the time.  That is the type of PTL I'd like to be to help
repay the community.
>
> 2. What are your plans to make the Cinder project as a core component
> better (no... really, what specifically and how does it make Cinder
> better)?
>
> Most candidates are representing a storage vendor naturally.  Everyone
> says "make Cinder better"; But how do you intend to balance vendor
> interest and the interest of the general project?  Where will your
> focus in the M release be?  On your vendor code or on Cinder as a
> whole?  Note; I'm not suggesting that anybody isn't doing the "right"
> thing here, I'm just asking for specifics.
  I believe I detailed some of these in my candidacy letter.   I firmly
believe that there are some Nova and Cinder interactions that need to
get fixed.  This will be a good first step along the way to allowing
active/active c-vol services.   Making Cinder better means not only
guiding the direction of features and fixes, but it also means
encouraging the community of driver developers to get involved and
informed about Cinder core itself.   We need a cinder driver developer
how to guide.  There are some items for driver developers that they need
to be aware of, and it would be great to be able to point folks to that
place.  For example, Fibre Channel drivers need to use the Fibre Channel
Zone Manager utils decorators during initialize_connection and
terminate_connection time.   Also, during terminate_connection time, a
driver needs to not always return the initiator_target_map.   Where is
that documented?  It's not, and it's only being caught in reviews.  The
trick as always is keeping that guide relevant with updates. 

   I've been pretty fortunate at HP, to be able to convince my
management, that working on Cinder specific issues as a priority, such
as multi-attach, os-brick, live migration, Nova <--> Cinder interactions
to name a few.   My team at HP isn't just responsible for maintaining
3PAR/LeftHand drivers to Cinder.  We are also involved in making Cinder
a more robust, scalable project, so that we can make a better Helion
product for our customers.  Helion is OpenStack and how we work on
Helion is to first and foremost work on OpenStack Cinder and Nova.   So,
from my perspective. HP's interests allow me to work on Cinder core
first and foremost. 

>
> 3. ​Why do you want to be PTL for Cinder?
>
> Seems like a silly question, but really when you start asking that
> question the answers can be surprising and somewhat enlightening. 
> There's different motivators for people, what's yours?  By the way,
> "my employer pays me a big bonus if I win" is a perfectly acceptable
> answer in my opinion, I'd prefer honesty over anything else.  You may
> not get my vote, but you'd get respect.
I've been working on various Open Source projects 

Re: [openstack-dev] [CINDER] [PTL Candidates] Questions

[Erlon Cruz]

John,

Thanks for the questions, it Ill really help me to make a the best choice.
I hadn't pondered the first question . +1 to make those and other if
suggest part of the candidate proposal.

Erlon

On Mon, Sep 21, 2015 at 6:49 AM, Duncan Thomas 
wrote:

> Hi John. Thanks for the questions.
>
>
>> 1. Do you actually have the time to spend to be PTL
>>
>
> I'm very much aware, and discussed with my management prior to standing,
> that being PTL is a pretty much full time job. I realise I'm somewhat
> limited by not being in a US time zone, however I'm pretty flexible with
> working hours, and already spend a few evening a week working US hours. I'd
> also like to use my time-zone shift as an advantage - I'm aware of how
> difficult it is for non-US contributors to get really involved in cinder
> due to our (generally very efficient) IRC-centric nature. I'd like to see
> if we can make better use of the tools we have for getting attention on
> bugs, features and reviews.
>
>
> 2. What are your plans to make the Cinder project as a core component
>> better (no... really, what specifically and how does it make Cinder better)?
>>
>
> My main worry with Cinder is that we're drifting away from the core vision
> of both Openstack and the original Cinder team - A really good cloud, with
> really good block storage, no matter the technology behind it. We've so
> many half-finished features, APIs that only work under limited
> circumstances and general development debt that is seriously hurting us
> going forward. The new features being proposed are getting more niche, more
> 'everything and the kitchen sink' and less 'top quality, rock solid
> service'. I'd like to shift a focus on back-to-basics, and work on fixing
> the road blocks to fixing these issues - we have plenty of competent
> motivated people, but communication and bureaucratic issues issues both
> within our team and between cinder and other projects (primarily but not
> limited to nova and glance) have gotten in the way.
>
> Things I'd like to see done this cycle:
> - Python3 work - let's just push through it and get it done. Maybe focus
> on it exclusively for a few days or a week some time this cycle. It's
> dragging on, and since we aren't at the point where cinder actually runs
> under python3, new problems slip in regularly.
>
> - Replication, CGs, online backup etc rolled out to more drivers. Lets
> limit the amount of new things drivers need to add this cycle until we've
> caught up on the backlog.
>
> - Nova <-> cinder API. Fixing this in a way that works for the nova team
> appears to need micro-versions. This API has been a thorn in our side for
> all sorts of new features and bugs many times, let's tame it.
>
> - Making CI failures easier to understand. I really struggle to read most
> CI failures, and so don't follow up on them as often as I should. I'm sure
> I'm not alone. I'm convinced that a small amount of work with white space,
> headings etc in devstack and tempest logs could give a really big boost.
> I'd also like to see a state other than 'failed' for situations where there
> was a problem with the CI system itself and so it didn't get as far as
> trying to deploy devstack. As I mentioned, we've enough smart people to
> make improvements that should allow us all to be more productive
>
> - Reducing review noise. I suspect that some policing and emailing people
> to improve etiquette on reviews (don't -1 for spelling and grammar, don't
> post a review until it is ready to be reviewed, give people time to batch
> comments rather than posting a new version for every nit, etc) will pay
> off, but it needs time dedicated to it.
>
> - Less out-of-band discussion on community decisions. I'm a big believer
> that discussion on record and in public, either on IRC or email, has much
> more value than private discussions and public statements. It also reduces
> accusations of bias and unfairness.
>
>
>> 3. ​Why do you want to be PTL for Cinder?
>>
>
> I wan to see cinder continue to succeed. My code contributions have, for
> various reasons, reduced in quantity and value against my efforts on
> mentoring, designs, reviews and communications. I'd like to free up the
> people who are actually writing good code to do more of that, by taking on
> more of the non-code burden and working to remove road blocks that are
> stopping people from making progress - be those internally with-in the
> team, between openstack teams or even helping people solve problems
> (managerial, legal or educational) within their own companies. I've had a
> fair bit of success at that in the past, and I believe that now is the time
> when those skills are the most effective ones to move cinder forward. We've
> a great technical team, so I want to enable them to do more, while keeping
> on top of scope creep and non-standardisation enough to enable cinder to be
> what I and many others would like it be.
>
>
>
>
> I hope this helps people with 

Re: [openstack-dev] [CINDER] [PTL Candidates] Questions

[Duncan Thomas]

Hi John. Thanks for the questions.


> 1. Do you actually have the time to spend to be PTL
>

I'm very much aware, and discussed with my management prior to standing,
that being PTL is a pretty much full time job. I realise I'm somewhat
limited by not being in a US time zone, however I'm pretty flexible with
working hours, and already spend a few evening a week working US hours. I'd
also like to use my time-zone shift as an advantage - I'm aware of how
difficult it is for non-US contributors to get really involved in cinder
due to our (generally very efficient) IRC-centric nature. I'd like to see
if we can make better use of the tools we have for getting attention on
bugs, features and reviews.


2. What are your plans to make the Cinder project as a core component
> better (no... really, what specifically and how does it make Cinder better)?
>

My main worry with Cinder is that we're drifting away from the core vision
of both Openstack and the original Cinder team - A really good cloud, with
really good block storage, no matter the technology behind it. We've so
many half-finished features, APIs that only work under limited
circumstances and general development debt that is seriously hurting us
going forward. The new features being proposed are getting more niche, more
'everything and the kitchen sink' and less 'top quality, rock solid
service'. I'd like to shift a focus on back-to-basics, and work on fixing
the road blocks to fixing these issues - we have plenty of competent
motivated people, but communication and bureaucratic issues issues both
within our team and between cinder and other projects (primarily but not
limited to nova and glance) have gotten in the way.

Things I'd like to see done this cycle:
- Python3 work - let's just push through it and get it done. Maybe focus on
it exclusively for a few days or a week some time this cycle. It's dragging
on, and since we aren't at the point where cinder actually runs under
python3, new problems slip in regularly.

- Replication, CGs, online backup etc rolled out to more drivers. Lets
limit the amount of new things drivers need to add this cycle until we've
caught up on the backlog.

- Nova <-> cinder API. Fixing this in a way that works for the nova team
appears to need micro-versions. This API has been a thorn in our side for
all sorts of new features and bugs many times, let's tame it.

- Making CI failures easier to understand. I really struggle to read most
CI failures, and so don't follow up on them as often as I should. I'm sure
I'm not alone. I'm convinced that a small amount of work with white space,
headings etc in devstack and tempest logs could give a really big boost.
I'd also like to see a state other than 'failed' for situations where there
was a problem with the CI system itself and so it didn't get as far as
trying to deploy devstack. As I mentioned, we've enough smart people to
make improvements that should allow us all to be more productive

- Reducing review noise. I suspect that some policing and emailing people
to improve etiquette on reviews (don't -1 for spelling and grammar, don't
post a review until it is ready to be reviewed, give people time to batch
comments rather than posting a new version for every nit, etc) will pay
off, but it needs time dedicated to it.

- Less out-of-band discussion on community decisions. I'm a big believer
that discussion on record and in public, either on IRC or email, has much
more value than private discussions and public statements. It also reduces
accusations of bias and unfairness.


> 3. ​Why do you want to be PTL for Cinder?
>

I wan to see cinder continue to succeed. My code contributions have, for
various reasons, reduced in quantity and value against my efforts on
mentoring, designs, reviews and communications. I'd like to free up the
people who are actually writing good code to do more of that, by taking on
more of the non-code burden and working to remove road blocks that are
stopping people from making progress - be those internally with-in the
team, between openstack teams or even helping people solve problems
(managerial, legal or educational) within their own companies. I've had a
fair bit of success at that in the past, and I believe that now is the time
when those skills are the most effective ones to move cinder forward. We've
a great technical team, so I want to enable them to do more, while keeping
on top of scope creep and non-standardisation enough to enable cinder to be
what I and many others would like it be.




I hope this helps people with their decision. Whomever wins, I have high
hopes for the future, there is nobody standing who hasn't been a pleasure
to work with in the past, and I don't expect that to change in the future.

-- 
Duncan Thomas
__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe

Re: [openstack-dev] [CINDER] [PTL Candidates] Questions

[Sean McGinnis]

On Sun, Sep 20, 2015 at 11:30:15AM -0600, John Griffith wrote:
> ​PTL nomination emails are good, but I have a few questions that I'd like
> to ask to help me in making my vote.  Some of these are covered in the
> general proposal announcements, but I'd love to hear some more detail.
> 
> It would be awesome if the Cinder candidates could spend some time and
> answer these to help me (and maybe others) make an informed choice:

Great idea John. We have a lot of candidates this time around, so it's
probably a good idea to get a little more info before the election is
over.

> 
> 1. Do you actually have the time to spend to be PTL
> 

Yes. Prior to submitting my name I had a few conversations with my
management to make sure this would be something they would support.

I have been assured I could make Cinder my primary and full time
responsibility should I become elected.

> 
> 2. What are your plans to make the Cinder project as a core component
> better (no... really, what specifically and how does it make Cinder better)?
> 
> Most candidates are representing a storage vendor naturally.  Everyone says
> "make Cinder better"; But how do you intend to balance vendor interest and
> the interest of the general project?  Where will your focus in the M
> release be?  On your vendor code or on Cinder as a whole?  Note; I'm not
> suggesting that anybody isn't doing the "right" thing here, I'm just asking
> for specifics.

Even though we have some vendor code in Cinder, I'm lucky enough to have
a couple folks on my team that I've been having take care of anything to
do with our driver. My role would be specifically to focus on the core
and overall (multi-vendor contributions, cross-project collaboration, etc.)
contributions.

I can't say I have a specific, actionable plan for exactly what I would
do to "make Cinder better". I think we already have several initiatives 
under way in that respect that I would help to make those happen. I
would see my role as more of a facilitator to help provide support and
focus resources on accomplishing these goals.

I would also work with OpenStack operators, other projects that are
consumers of Cinder services, and the community at large to make sure
Cinder is meeting their block storage needs.

> 
> 3. ​Why do you want to be PTL for Cinder?
> 
> Seems like a silly question, but really when you start asking that question
> the answers can be surprising and somewhat enlightening.  There's different
> motivators for people, what's yours?  By the way, "my employer pays me a
> big bonus if I win" is a perfectly acceptable answer in my opinion, I'd
> prefer honesty over anything else.  You may not get my vote, but you'd get
> respect.

I won't get a big bonus, and I doubt I would get any kind of promotion
or increase out of this. What I would get is the ability to focus full
time on OpenStack and Cinder. Right now it is one of several of my 
responsibilities, and something that I spend a lot of my own time
on because I enjoy working on the project, and working with the folks
involved, and I believe in the future of OpenStack. 

I want to be PTL because I feel I could be a "facilitator" of all the
different efforts underway, to help drive them to completion and to
help reduce the distractions away from all the smart folks that are
getting things done. I can organize, communicate, and simplify our efforts.

> 
> Thanks,
> John


__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Re: [openstack-dev] [CINDER] [PTL Candidates] Questions

[Ivan Kolodyazhny]

Hi John,

Thank you for these question. Such questions with answers could be a good
part of PTL proposal in the future.

Please, see my answers inline.

Regards,
Ivan Kolodyazhny

On Sun, Sep 20, 2015 at 8:30 PM, John Griffith 
wrote:

> ​PTL nomination emails are good, but I have a few questions that I'd like
> to ask to help me in making my vote.  Some of these are covered in the
> general proposal announcements, but I'd love to hear some more detail.
>
> It would be awesome if the Cinder candidates could spend some time and
> answer these to help me (and maybe others) make an informed choice:
>
> 1. Do you actually have the time to spend to be PTL
>
> I don't think many people realize the time commitment. Between being on
> top of reviews and having a pretty consistent view of what's going on and
> in process; to meetings, questions on IRC, program management type stuff
> etc.
>

I sincerely admire if any PTL could be in TOP of reviews, commits, etc.
Cross-projects meetings and activities will take a lot of time. IRC
participation is required for every active contributor, especially for
PTLs. Talking about Cinder, we need to remember that not only Community is
involved into the project. Many vendors have their drivers and didn't
contribute to other parts of projects. Cinder PTL is also responsible for
communication and collaboration with vendors to make their drivers be
working with Cinder.


> Do you feel you'll have the ability for PTL to be your FULL Time job?
>
It was first question which I asked myself before a nomination.


> Don't forget you're working with folks in a community that spans multiple
> time zones.
>

Sure, I can't forget it because I spend time almost every night in our
#openstack-cinder channel. Talking about something more measurable, I would
like to provide only this http://stackalytics.com/report/users/e0ne report.



>
> 2. What are your plans to make the Cinder project as a core component
> better (no... really, what specifically and how does it make Cinder better)?
>
> Most candidates are representing a storage vendor naturally.  Everyone
> says "make Cinder better"; But how do you intend to balance vendor interest
> and the interest of the general project?  Where will your focus in the M
> release be?  On your vendor code or on Cinder as a whole?  Note; I'm not
> suggesting that anybody isn't doing the "right" thing here, I'm just asking
> for specifics.
>

My company doesn't have own driver. I don't want to talk about Block Device
Driver now. I'll be that person, who will create patch with removing after
M-2 milestone is this driver won't have CI and minimum required feature
set.

A a Cinder user and contributor, I'm interesting in making Cinder Core more
flexible (E.g. working w/o Nova), tested (e.g. functional tests, 3rd party
CI, unit tests coverage etc) and make our users more happy with it.


>
> 3. ​Why do you want to be PTL for Cinder?
>
> Seems like a silly question, but really when you start asking that
> question the answers can be surprising and somewhat enlightening.  There's
> different motivators for people, what's yours?  By the way, "my employer
> pays me a big bonus if I win" is a perfectly acceptable answer in my
> opinion, I'd prefer honesty over anything else.  You may not get my vote,
> but you'd get respect.
>

OpenStack itself is very dynamic project. It makes a big progress each
release. It varies greatly from one release to another. It's a real
community-driven project. And I think that   PTLs are not dictators. PTLs
only help community to work on Cinder to make it better. Each person as a
PTL could bring something new to the community. Sometimes, "something new"
may means "something bad", but after each mistake we'll do our work better.
Being PTL helps to understand not only Cinder developer needs. It helps to
understand other OpenStack project needs. PTL should take care not only on
Cider or Nova or Heat. IMO, The main task of each PTL is to coordinate
developers of one project, other OpenStack developers and vendors to work
together on regular basis. It will be very big challenge for me I'll do my
best to make Cinder better as PTL. I'm sure, Cinder community will help our
new PTL a lot with it.


> Thanks,
> John
>
> __
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
>

Ivan.
__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev