Re: [openstack-dev] [nova] bug discussion at mid cycle meet up

2014-07-30 Thread Daniel P. Berrange
On Tue, Jul 29, 2014 at 02:48:12PM -0400, Russell Bryant wrote:
 On 07/29/2014 11:43 AM, Tracy Jones wrote:
  3.  We have bugs that are really not bugs but features, or performance
  issues.  They really should be a BP not a bug, but we don’t want these
  things to fall off the radar so they are bugs… But we don’t really know
  what to do with them.  Should they be closed?  Should they have a
  different category – like feature request??  Perhaps they should just be
  wish list??
 
 I don't think blueprint are appropriate for tracking requests.  They
 should only be created when someone is proposing actually doing the work.

Agreed, we really do not want to see blueprints created for people
who are just doing 'drive by' feature requests.

 I think Wishlist is fine for keeping a list of requests.  That's what
 I've been using it for.

Yep, wishlist works fine IMHO - it is a nice low overhead approach
for users wanting to report these kind of items, with a tool they
are already familiar with.

Regards,
Daniel
-- 
|: http://berrange.com  -o-http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org  -o- http://virt-manager.org :|
|: http://autobuild.org   -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org   -o-   http://live.gnome.org/gtk-vnc :|

___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


[openstack-dev] [nova] bug discussion at mid cycle meet up

2014-07-29 Thread Tracy Jones

At the mid-cycle meet-up yesterday we spent some time looking at our bug 
dashboard (http://54.201.139.117/nova-bugs.html) and talking about things we 
can do to help focus on bugs.  We came up with the following ideas.  I’d like 
folks to weigh in on these i if you have some ideas or concerns.


1.  Start auto-abandoning bugs that have not been touched (I.e. Updated) in the 
last 60 days.We would have something (a bot?) that would look at bugs that 
have not been updated (nor their review updated) in the last 60 days.  The bug 
would be set to the “new” state and the assignee would be removed.  This would 
cause the bug to be re-triaged and would be up for someone else to pick up.

2.  Also - when a bug has all abandoned reviews, we should automatically set 
the bug to new and remove the assignee.

3.  We have bugs that are really not bugs but features, or performance issues.  
They really should be a BP not a bug, but we don’t want these things to fall 
off the radar so they are bugs… But we don’t really know what to do with them.  
Should they be closed?  Should they have a different category – like feature 
request??  Perhaps they should just be wish list??

4.  We should  have more frequent and focused bug days.  For example every 
Monday have a bug day where we focus on 1 area (like api or compute or 
networking for example) and work on moving new bugs to configured, or confirmed 
bugs to triaged.  I’ll talk with Michael about when to schedule the 1st 
“focused” bug day and what area to address.


In generate we need to tighten up the definition of triaged and confirmed.  
Bugs should move from New - Confirmed - Triaged - In Progress.  JayPipes has 
updated the wiki to clarify this.

  *   Confirmed means someone has looked at the bug, saw there was enough into 
to start to diagnose, and agreed it sounds like a bug.
  *   Triaged means someone has analyzed the bug and can propose a solution 
(not necessarily a patch).  If the person is not going to fix it, they should 
update the bug with the proposal and move the bug into Triaged.

If we do implement 1 and 2 I’m hoping the infra team can help – I think dims 
volunteered ;-)

I made a note to add review stats to the bug page to make it easier to see how 
far along a bug is in review




___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


Re: [openstack-dev] [nova] bug discussion at mid cycle meet up

2014-07-29 Thread Russell Bryant
On 07/29/2014 11:43 AM, Tracy Jones wrote:
 3.  We have bugs that are really not bugs but features, or performance
 issues.  They really should be a BP not a bug, but we don’t want these
 things to fall off the radar so they are bugs… But we don’t really know
 what to do with them.  Should they be closed?  Should they have a
 different category – like feature request??  Perhaps they should just be
 wish list??

I don't think blueprint are appropriate for tracking requests.  They
should only be created when someone is proposing actually doing the work.

I think Wishlist is fine for keeping a list of requests.  That's what
I've been using it for.

 In generate we need to tighten up the definition of triaged and
 confirmed.  Bugs should move from New - Confirmed - Triaged - In
 Progress.  JayPipes has updated the wiki to clarify this.
 
   * Confirmed means someone has looked at the bug, saw there was enough
 into to start to diagnose, and agreed it sounds like a bug.  
   * Triaged means someone has analyzed the bug and can propose a
 solution (not necessarily a patch).  If the person is not going to
 fix it, they should update the bug with the proposal and move the
 bug into Triaged.  

We should be careful not to conflict with the guidelines set for all
OpenStack projects here:

   https://wiki.openstack.org/wiki/BugTriage

For example, that page says when a bug should be set to Confirmed or
Triaged.  In most cases, it's Confirmed.  Triage is when there is a
known solution.

-- 
Russell Bryant

___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


Re: [openstack-dev] [nova] bug discussion at mid cycle meet up

2014-07-29 Thread Jay Pipes

On 07/29/2014 11:48 AM, Russell Bryant wrote:

On 07/29/2014 11:43 AM, Tracy Jones wrote:

3.  We have bugs that are really not bugs but features, or performance
issues.  They really should be a BP not a bug, but we don’t want these
things to fall off the radar so they are bugs… But we don’t really know
what to do with them.  Should they be closed?  Should they have a
different category – like feature request??  Perhaps they should just be
wish list??


I don't think blueprint are appropriate for tracking requests.  They
should only be created when someone is proposing actually doing the work.

I think Wishlist is fine for keeping a list of requests.  That's what
I've been using it for.


There's a metric crap-ton of bugs that are *not* in Wishlist and are 
instead in High or Medium importance, but they are not necessarily bugs 
that either have a specific solution -- see: performance-related bugs -- 
or are things that frankly can never be fixed.


We don't want to keep these things as bugs, because they aren't really 
bugs, but the blueprint/spec stuff isn't appropriate for epics that 
are like super-specs to be used to track general themes. We think that a 
separate category of thing is needed for tracking these themes. In 
Agile, these things are called epics.



In generate we need to tighten up the definition of triaged and
confirmed.  Bugs should move from New - Confirmed - Triaged - In
Progress.  JayPipes has updated the wiki to clarify this.

   * Confirmed means someone has looked at the bug, saw there was enough
 into to start to diagnose, and agreed it sounds like a bug.
   * Triaged means someone has analyzed the bug and can propose a
 solution (not necessarily a patch).  If the person is not going to
 fix it, they should update the bug with the proposal and move the
 bug into Triaged.


We should be careful not to conflict with the guidelines set for all
OpenStack projects here:

https://wiki.openstack.org/wiki/BugTriage

For example, that page says when a bug should be set to Confirmed or
Triaged.  In most cases, it's Confirmed.  Triage is when there is a
known solution.


So, yeah, we went back and forth on this. One thing that was mentioned 
is that by setting something to Wishlist from, say, High, we downplay 
the importance of the particular bug, which, for performance and 
scalability epics tends to annoy both the bug submitter and the bug owner.


However, setting the bug to New, which triggers a re-verification and/or 
re-triaging of the issue, puts the onus and responsibility on the bug 
triaging team and allows the bug to be taken off of the In progress but 
Abandoned list and not lost into the general swamp of In progress but 
not assigned.


Anyway, just food for thought.
-jay


___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev