Description of problem: ====================== When a brick of a EC cold layer is down, promotions fail due to following error [2015-10-16 06:43:59.589013] E [socket.c:2278:socket_connect_finish] 0-ecvol-client-5: connection to 10.70.34.43:49243 failed (Connection refused) [2015-10-16 06:44:00.128952] E [MSGID: 109037] [tier.c:939:tier_process_brick] 0-tier: Failed gettingjournal_mode of sql db /rhs/brick3/ecvol/.glusterfs/ecvol.db [2015-10-16 06:44:00.128989] E [MSGID: 109087] [tier.c:1033:tier_build_migration_qfile] 0-ecvol-tier-dht: Brick query failed While the error makes complete sense, we must handle this scenario, where db can be built or vetoed using the remaining up and running bricks as long as EC quorum is not breached. I.e in a 4+2 ec cold layer, even if 2 bricks are down, then the data is as good as available as the quorum of 4 bricks is met. hence db should be built and used based on these 4. Version-Release number of selected component (if applicable): ============================================================ glusterfs-server-3.7.5-0.22.gitb8ba012.el7.centos.x86_64 Steps to Reproduce: ================= 1.create a ec 4+2 vol and start it 2.Now create a file f1 and then bring down 1 or 2 bricks 3.Now modify f1 4. attach tier and modify f1 to heat it 5. Now create new file h1 and wait for it to be demoted 6. Now touch/modify h1 after it is demoted Actual results: =================== both f1 and h1 fail to get promoted
logs and sosreport available @ /home/repo/sosreports/nchilaka/bug.1272334 [nchilaka@rhsqe-repo bug.1272334]$ hostname rhsqe-repo.lab.eng.blr.redhat.com
REVIEW: http://review.gluster.org/12405 (cluster/tier do not abort migration if a single brick is down) posted (#1) for review on release-3.7 by Dan Lambright (dlambrig)
COMMIT: http://review.gluster.org/12405 committed in release-3.7 by Dan Lambright (dlambrig) ------ commit 6c6b4bb361fb6fa3adc69e43d185c755b2f4c771 Author: Dan Lambright <dlambrig> Date: Mon Oct 19 20:42:56 2015 -0400 cluster/tier do not abort migration if a single brick is down backport fix 12397 When a bricks are down, promotion/demotion should still be possible. For example, if an EC brick is down, the other bricks are able to recover the data and migrate it. > Change-Id: I8e650c640bce22a3ad23d75c363fbb9fd027d705 > BUG: 1273215 > Signed-off-by: Dan Lambright <dlambrig> > Reviewed-on: http://review.gluster.org/12397 > Tested-by: NetBSD Build System <jenkins.org> > Tested-by: Gluster Build System <jenkins.com> > Reviewed-by: Joseph Fernandes Signed-off-by: Dan Lambright <dlambrig> Change-Id: I6688757eaf97426c8e1ea1038c598b34bf6b8ccc BUG: 1272334 Signed-off-by: Dan Lambright <dlambrig> Reviewed-on: http://review.gluster.org/12405 Tested-by: NetBSD Build System <jenkins.org> 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-3.7.6, please open a new bug report. glusterfs-3.7.6 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/gluster-users/2015-November/024359.html [2] http://thread.gmane.org/gmane.comp.file-systems.gluster.user