Re: avoiding incomplete backports

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, 2018-07-26 at 13:59 +0000, Sage Weil wrote:
> 
> These tend to make the process harder/slower but don't really get at
> the 
> part where someone says "should we be doing this?".  Two process-y
> options 
> come to mind:
> 
>  1- In the 'Pending backport' state, the backport bugs get a
> 'Pending 
> approval' state.  Or, 'pending backport approval' come sbefore
> 'pending 
> backport'.  Or, a new 'backport approved' checkbox field for
> backports.  
> As long as it's something that we can build a redmine query for.  And
> the 
> team lead or a core maintainer has to say yes or no before the 
> backporting work begins.
>  2- Once the backport PR is opened, a review is always requested 
> from the lead.
> 
> Or,
> 
>  3- We collectively pay more attention to the reviews for backports
> and 
> consider whether the backport is really a good idea and worth the
> risk.
> 

Well, backporting will always be a tradeoff between time spent handling
them (the code, the process), and safety/correctness.

While 3 relaxes the process by making it simpler, it will also impact
the confidence we'll have in them. This will not be a huge problem when
dealing with small, trivial backports, but for anything besides that,
this can easily come back to bite us.

That said, I'm very much in favor of 1/2, although that doesn't mean 3
can't fit in.

IMO, a simple process would be something like:

1. everything requires a ticket in the tracker, ideally referencing the
original ticket.
2. the checkbox you mentioned, "backport approved" indicates
backporting is okay, and should come only after the code has been
reviewed by a lead/core maintainer (which likely requires them being
asked for review on PRs).
3. anyone is welcome to review backports, and provide their feedback on
the tracker ticket; it will be upon the leads to take that into
consideration.

I think 3. may simplify the leads' life, if we collectively help out
ascertaining whether something is valuable, safe, correct. The thing
that strikes me as annoying is the checkbox in 2. For it to be
feasible, it would likely require multiple tickets, for each release
being backported to, for the checkbox to make sense. OTOH, it would
make sense to have multiple tickets (one per release) anyway.

  -Joao

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html



[Index of Archives]     [CEPH Users]     [Ceph Large]     [Information on CEPH]     [Linux BTRFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux