[heketi-devel] Meeting Minutes for 2017.12.20

Jose A. Rivera jarrpa at redhat.com
Tue Dec 20 16:21:50 UTC 2016


Discussions on Release 4:

   -

   Tentative release date for January 21, 2017
   -

   Need to overhaul the centos-ci setup process
   -

      Use prebuilt boxes: https://github.com/heketi/heketi/issues/629
      -

      Improve docker image build (e.g. building in a vm or container in
      order to avoid the dvm workaround…)
      -

      Host scripts and instructions for creating centos-ci boxes on
      https://github.com/heketi/vagrant-heketi or
      https://github.com/gluster/glusterfs-patch-acceptance-tests/
      -

      Use pre-existing VM for building the docker image
      -

   Going through all the Issues attached to Release 4, deciding whether
   we're taking or deferring them
   -

   Discussion on minimum brick size in heketi:
   -

      Currently 4GB, some would like to be able to make it 1GB
      -

      Raise an issue for release 4: Change minimum brick size to 1GB
      -

         Claimed by Ashiq
         -

         Make sure to update (or add) unit tests!
         -

      Raise an issue for release 4: Change size of db-storage volume to
      smallest possible (1GB?)
      -

         Claimed by Ashiq
         -

         Make sure to update (or add) unit tests!


Discussion on future releases:

   -

   Release 5: Tentative date for April, 2017
   -

   Subsequent releases: Every 3-6 months


Future issues to discuss:

   -

   Tightening integration with GlusterFS
   -

      Might remove the need for most of the heketi DB
      -

      Needs to be fast on the GlusterFS side (e.g. REST API)
      -

      Likely for glusterd2
      -

   Improve user experiences:
   -

      What happens when a disk or node dies?
      -

         https://github.com/heketi/heketi/issues/161
         -

         Automatically create new brick(s) to preserve replica, rebalance
         -

         Implement "disk fail" REST command
         -

         Would allow the ability to replace/swap out disks before or when
         they fail
         -

      Volume snapshot management
      -

         https://github.com/heketi/heketi/issues/413
         -

         Need to do :)

Meeting doc:
https://docs.google.com/document/d/1oy0x-gG8S1-950rGG8KFHdFiY8WSFHhGz5OMlKHmX0E/edit
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/heketi-devel/attachments/20161220/437ee85f/attachment-0001.html>


More information about the heketi-devel mailing list