Description of problem: Peer goes to rejected state after peer probing a new 3.7.x node from 3.6.x when quota is enabled. 1) Create a volume(a plain distribute volume with one brick) with gluster 3.6.1 2) Enable quota(Check the quota.conf file and it is v1.1) 3) Now upgrade the gluster version to 3.7.1-16 4) Bump up the op-version using 'gluster v set all cluster.op-version 30703' 5) The version in quota.conf is still v1.1. Now do a peer probe <node1>. Peer probe will succeed but the new node(node1) now has a quota version v1.2 while the node from which peer probe is initiated has still quota version as v1.1 in quota.conf. 6) Issue gluster peer status. The new node(node1) will now has the status as Peer rejected.
REVIEW: http://review.gluster.org/15791 (glusterd/quota: upgrade quota.conf file during an upgrade) posted (#2) for review on release-3.8 by sanoj-unnikrishnan (sunnikri)
COMMIT: http://review.gluster.org/15791 committed in release-3.8 by Atin Mukherjee (amukherj) ------ commit 9580864838c7c6044368cd969d1c11a2aad37c3c Author: Manikandan Selvaganesh <mselvaga> Date: Tue Aug 30 17:53:09 2016 +0530 glusterd/quota: upgrade quota.conf file during an upgrade Problem ======= When quota is enabled on 3.6, it will have quota conf version in quota.conf as v1.1. This node gets upgraded to 3.7 but it will still have quota conf version as v1.1 until a quota enable/disable/set limit is initiated. When this is not initiated and when this node tries to peer probe a node which is a fresh install of 3.7 (which will have quota conf version as v1.2), then this will result in "Peer rejected" state. This patch fixes the issue. Solution ======== When an upgrade happens from 3.6 to 3.7, quota.conf file needs to be modified as well. With 3.6, in quota.conf the version will be v1.1 and it needs to be changed to v1.2 from 3.7. This is because in 3.7, inode quota feature is introduced. So when an op-version bumpup happens quota.conf needs to be upgraded with quota conf version v1.2 and all the 16 byte uuid needs to be changed to 17 bytes uuid as well. Previously, when the cluster version is upgraded to 3.7, the quota.conf got upgraded as well. But, the upgradation was done only when quota enable/disable/set limit is done. With this patch, the upgradation is done during a cluster op version bump up as well. > Reviewed-on: http://review.gluster.org/15352 > Tested-by: Atin Mukherjee <amukherj> > NetBSD-regression: NetBSD Build System <jenkins.org> > CentOS-regression: Gluster Build System <jenkins.org> > Smoke: Gluster Build System <jenkins.org> > Reviewed-by: Atin Mukherjee <amukherj> (cherry picked from commit 4b2cff614462508eef529c5d128e0974720e3f50) Change-Id: Idb5ba29d3e1ea0e45c85d87c952c75da9e0f99f0 BUG: 1392716 Signed-off-by: Manikandan Selvaganesh <mselvaga> Reviewed-on: http://review.gluster.org/15791 Smoke: Gluster Build System <jenkins.org> Reviewed-by: Manikandan Selvaganesh <manikandancs333> Tested-by: Manikandan Selvaganesh <manikandancs333> NetBSD-regression: NetBSD Build System <jenkins.org> CentOS-regression: Gluster Build System <jenkins.org>
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-3.8.6, please open a new bug report. glusterfs-3.8.6 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://www.gluster.org/pipermail/packaging/2016-November/000217.html [2] https://www.gluster.org/pipermail/gluster-users/