[Gluster-users] Plans for Gluster 3.8

Vijay Bellur vbellur at redhat.com
Fri Jul 10 03:07:06 UTC 2015

On Friday 10 July 2015 03:22 AM, John Gardeniers wrote:
> Vijay,
> Instead of thinking about the next version, shouldn't you be trying to
> get at least one truly stable version? 3.5 still has problems and the
> only things I've been reading about 3.6 and 3.7 are problems, so they
> are far from stable. This idea of releasing a new version instead of
> fixing the current one is merely driving people away from Gluster.

Our intention is to do both - have stability releases happening in the 
3.5.x, 3.6.x and 3.7.x trains while we develop the next Gluster releases.

If we look at the previous 6-12 month period, there has been at least 
one minor release per month as we released 3.6 and 3.7. If the minor 
releases are not being very effective in addressing stability, we would 
love to have some inputs on why that's not happening. Something like:

a> Minor releases do not have the content/bug fixes that you are looking for

b> Minor releases should happen even more frequently

c> Fix for my issue is present in master but has not been backported to 
a particular release train

and more along these lines would surely help us in doing a course 

OTOH, we are *extremely* serious about stability and there is an ever 
increasing focus on testing & feedback from testing. Some of the recent 
actions in this space include:

- Weekly IRC bug triage meetings for prioritizing bugs

- Addition of dedicated maintainers for tests in gluster

- More pre-commit & post-commit tests being integrated with Jenkins

Thanks for bringing this up. Feedback of this nature with more details 
will help us determine the content for 3.8. If there are issues that you 
would like to be addressed, please let us know. We can either fix them 
in the appropriate release trains if the nature of changes is not 
significant or if it happens to be so, we can address the issues as 
enhancements in 3.8.


More information about the Gluster-users mailing list