[Gluster-devel] Reviews needed before glusterfs-3.5.1 beta2 can get released

[Niels de Vos]

Hopefully we can get a beta2 for glusterfs-3.5.1 out in the following 
days. Most of the outstanding patches have now passed regression testing 
and are merged. There are still a few bugs listed as blockers, and their 
patches need to be included in the master branch before I'll merge their 
counter parts in release-3.5.

Please review these patches as soon as possible. In case corrections are 
required, check with the owner of the change and try to help posting an 
updated version if possible.

- http://review.gluster.org/7963 (glusterd: Better op-version values and 
  ranges)
  - http://review.gluster.org/8010 for release-3.5

- http://review.gluster.org/8004 (gNFS: Make NFS DRC off by default)
  - no patch posted for release-3.5, still needs to be done

- http://review.gluster.org/7853 (doc: Fix peer probe deficiencies)
  - missing matching patch for the master branch?

The only other bug not in MODIFIED or ON_QA in the list of blockers [1] 
is related to the cppcheck auditing. Lala and Kaleb have done some work 
on this and I have contacted them for getting the current status. We'll 
see if these changes can make it in 3.5.1, or get scheduled for 3.5.2.

After 3.5.1 beta2 has been released, we need some verifications that its 
not horribly broken. The only reported issue with the 1st beta was 
related to the op-version and server.manage-gids. This should be 
resolved with the first two patches earlier in this email. Some other 
issues that are in 3.5.0 have been addressed in addition, making 3.5.1 
an important update (list of changes below).

Thanks,
Niels

[1] https://bugzilla.redhat.com/showdependencytree.cgi?id=glusterfs-3.5.1&hide_resolved=1

v3.5.1beta to current release-3.5/HEAD:
* features/gfid-access: calloc gfid and set in xdata
* glusterd: On gaining quorum spawn_daemons in new thread
* NetBSD build fixes for release-3.5
* logging: use duplicate stderr, instead of re-opening
* heal: Don't depend on errno for reporting heal info status
* doc: Document libgfapi and its apis,usage,bindings..etc
* doc: Add 'expose volume capabilities' details of bd-xlator
* features/glupy: GPLv2 or LGPLv3+ license
* doc/rdmacm: fix formatting errors.
* doc: add readdir-ahead documentation
* doc: fix brick port ranges in admin guide.
* doc: update details about posix acl on nfs mount
* doc: Add details on qemu glusterfs integration
* feature/gfid-access: Adding documentation
* doc: quota-scalability documentation
* transport/rdma: Add rdma connection manager documentation
* doc:Adding documentation for glusterfs and ovirt integration

v3.5.0 to v3.5.1beta:
* logging: remove unused message-id scripts
* doc: distributed geo-replication
* tests/rpm: always run ./autogen.sh to create missing files
* doc: changelog consumer library
* rpc: implement server.manage-gids for group resolving on the bricks
* rpc: warn and truncate grouplist if RPC/AUTH can not hold everything
* libgfapi: Added Handle-based ops to get/set/remove extended attributes in the libgfapi.
* doc:  Add block device xlator details
* doc: zerofill api documentation
* doc: afr-statistics command  documentation
* cluster/afr: remove unused variable
* cluster/afr: send opendirs to all children for entry self-heal
* cluster/afr: Remove eager-lock stub on finodelk failure
* libgfapi:  Added support to fetch volume info from glusterd and store in glfs object.
* doc: Adding user documentation for WORM feature
* Removed the old and unused non-Glupy Python bindings code
* build: glusterfs.spec.in, minor/nit changes to sync with Fedora spec
* features/marker-quota: more stringent error handling in rename.
* cluster/afr: Fix bugs in quorum implementation
* gfapi: glfs_set_volfile_server() now entertains multiple calls
* posix: if brick-uid or brick-gid is not specified, do not set
* write-behind: track filesize when doing extending writes
* fuse: prevent READDIR(P) from writing to much data to /dev/fuse
* cluster/afr: Unable to self heal symbolic links
* features/locks: Remove stale entrylk objects from 'blocked_locks' list
* rpcsvc: Ignore INODELK/ENTRYLK/LK for throttling
* storage/posix:  lgetxattr called with invalid keys on the bricks
* storage/posix: do not dereference gfid symlinks before posix_handle_mkdir_hashes()
* build: do not create versioned <xlator>.so files
* cluster/afr: perform list-xattr during lookup
* features/locks: Remove stale inodelk objects from 'blocked_locks' list
* storage/posix: add list xattr capability to lookup
* cluster/afr: Prevent heal info hang when data-self-heal in progress.
* heal: Enable logging for glfsheal.
* cluster/afr: trigger self-heals even when they are set to off.
* cluster/afr: get virtual-xattrs only on valid xdata.
* Tests: Changes to tests for glfs-heal implementation.
* heal: glfs-heal implementation
* cluster/afr: Add dry-run functionality to self-heal.
* cluster/afr: Add foreground self-heal launch capability through lookup.
* doc: Add glfs_fini known issue to release notes 3.5
* doc: Adding documentation for file-snapshot
* rdma: correct some spelling mistakes
* doc: Add doc for network compression xlator
* Release-notes for release-3.5