From e6045103b7e010779549bb486c00a07b3c3eb0fc Mon Sep 17 00:00:00 2001 From: Jeff Darcy Date: Mon, 20 Mar 2017 12:32:33 -0400 Subject: glusterd: hold off volume deletes while still restarting bricks We need to do this because modifying the volume/brick tree while glusterd_restart_bricks is still walking it can lead to segfaults. Without waiting we could accidentally "slip in" while attach_brick has released big_lock between retries and make such a modification. Backport of: > Commit a7ce0548b7969050644891cd90c0bf134fa1594c > BUG: 1432542 > Reviewed-on: https://review.gluster.org/16927 Change-Id: I30ccc4efa8d286aae847250f5d4fb28956a74b03 BUG: 1441476 Signed-off-by: Jeff Darcy Reviewed-on: https://review.gluster.org/17044 NetBSD-regression: NetBSD Build System Smoke: Gluster Build System CentOS-regression: Gluster Build System Reviewed-by: Atin Mukherjee --- xlators/mgmt/glusterd/src/glusterd.h | 1 + 1 file changed, 1 insertion(+) (limited to 'xlators/mgmt/glusterd/src/glusterd.h') diff --git a/xlators/mgmt/glusterd/src/glusterd.h b/xlators/mgmt/glusterd/src/glusterd.h index f3c7e1d6891..b99d433db74 100644 --- a/xlators/mgmt/glusterd/src/glusterd.h +++ b/xlators/mgmt/glusterd/src/glusterd.h @@ -192,6 +192,7 @@ typedef struct { int ping_timeout; uint32_t generation; int32_t workers; + uint32_t blockers; } glusterd_conf_t; -- cgit