Description of problem: Volume-snapshot feature inherits the notion of quorum from the server-side quorum feature. In this feature, quorum is computed on the set of peers participating in the ongoing transaction. This is incorrect. For e.g, Snapshot create should fail if there are less than 50% (by default) of nodes are online in a cluster. Let's take a storage pool made of 5 nodes. We need at least 3 nodes online for snapshot create command to succeed. But if we computed quorum on the no. of nodes participating (i.e online at the beginning of the transaction) then we would violate the pre-requisite for successful completion of snapshot-create. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. Create a volume and start it. 2. Bring a few nodes offline, such that server-quorum* (50% default) is not met. 3. Take a volume-snapshot using snapshot-create command. * - Server-quorum need not be configured for snapshot-create to function properly. Actual results: Snapshot create will succeed. Expected results: Snapshot create should fail. Additional info:
REVIEW: http://review.gluster.org/9493 (glusterd: compute quorum on peers in cluster) posted (#4) for review on master by Krishnan Parthasarathi (kparthas)
REVIEW: http://review.gluster.org/9492 (glusterd: group server-quorum related code together) posted (#4) for review on master by Krishnan Parthasarathi (kparthas)
COMMIT: http://review.gluster.org/9492 committed in master by Kaushal M (kaushal) ------ commit 4b18fba4064992e00cd5ebe1831afc79beab17b2 Author: Krishnan Parthasarathi <kparthas> Date: Tue Jan 27 15:12:03 2015 +0530 glusterd: group server-quorum related code together Server-quorum implementation was spread in many files. This patch brings them all together into a single file, namely glusterd-server-quorum.c. All exported functions are available via glusterd-server-quorum.h Change-Id: I8fd77114b5bc6b05127cb8a6a641e0295f0be7bb BUG: 1205592 Signed-off-by: Krishnan Parthasarathi <kparthas> Reviewed-on: http://review.gluster.org/9492 Reviewed-by: Atin Mukherjee <amukherj> Tested-by: Gluster Build System <jenkins.com> Reviewed-by: Kaushal M <kaushal>
REVIEW: http://review.gluster.org/10105 (mgmt/glusterd: set right definition of does_gd_meet_server_quorum()) posted (#1) for review on master by Vijay Bellur (vbellur)
REVIEW: http://review.gluster.org/10105 (glusterd: Fix rebase errors introduced by 4b18fba) posted (#2) for review on master by Kaushal M (kaushal)
COMMIT: http://review.gluster.org/10105 committed in master by Vijay Bellur (vbellur) ------ commit 12e603bb9d8bc220eeea01c0ecc8ca63f9df4bc4 Author: Vijay Bellur <vbellur> Date: Thu Apr 2 00:29:59 2015 +0530 glusterd: Fix rebase errors introduced by 4b18fba 4b18fba 'glusterd: group server-quorum related code together' a refactoring change, introduced some errors related to usage of userspace-rcu lists when rebasing. This patch fixes it. Changes done include, - Redo changes done by 673ba26 glusterd: Replace libglusterfs lists with liburcu lists - Redo changes done by c7785f7 glusterd: Protect the peer list and peerinfos with RCU. - Redo changes done by 891c7d0 glusterd: Prevent possible deadlock due to glusterd_quorum_count Change-Id: I789e5bc8b209d9ed6dd951d609baa90e89817639 BUG: 1205592 Signed-off-by: Vijay Bellur <vbellur> Signed-off-by: Kaushal M <kaushal> Reviewed-on: http://review.gluster.org/10105 Reviewed-by: Atin Mukherjee <amukherj> Reviewed-by: Niels de Vos <ndevos> Tested-by: Gluster Build System <jenkins.com>
REVIEW: http://review.gluster.org/9493 (glusterd: compute quorum on peers in cluster) posted (#5) for review on master by Kaushal M (kaushal)
COMMIT: http://review.gluster.org/9493 committed in master by Kaushal M (kaushal) ------ commit f66a85a484d41e68611f68d70e458c67f484d2e1 Author: Krishnan Parthasarathi <kparthas> Date: Tue Jan 27 16:44:20 2015 +0530 glusterd: compute quorum on peers in cluster ... and not on peers participating in an ongoing transaction. Change-Id: I6bdb80fd3bf3e7593fdf37e45a441d4a490469b8 BUG: 1205592 Signed-off-by: Kaushal M <kaushal> Reviewed-on: http://review.gluster.org/9493 Reviewed-by: Atin Mukherjee <amukherj> Tested-by: Gluster Build System <jenkins.com>
This bug is getting closed because a release has been made available that should address the reported issue. In case the problem is still not fixed with glusterfs-3.7.0, please open a new bug report. glusterfs-3.7.0 has been announced on the Gluster mailinglists [1], packages for several distributions should become available in the near future. Keep an eye on the Gluster Users mailinglist [2] and the update infrastructure for your distribution. [1] http://thread.gmane.org/gmane.comp.file-systems.gluster.devel/10939 [2] http://thread.gmane.org/gmane.comp.file-systems.gluster.user
*** Bug 1184393 has been marked as a duplicate of this bug. ***