From 5a7c86e578f5bbd793126a035c30e6b052177a9f Mon Sep 17 00:00:00 2001 From: Ravishankar N Date: Fri, 23 Dec 2016 07:11:13 +0000 Subject: afr: use accused matrix instead of readable matrix for deciding heals Problem: afr_replies_interpret() used the 'readable' matrix to trigger client side heals after inode refresh. But for arbiter, readable is always zero. So when `dd` is run with a data brick down, spurious data heals are are triggered. These heals open an fd, causing eager lock to be disabled (open fd count >1) in afr transactions, leading to extra FXATTROPS Fix: Use the accused matrix (derived from interpreting the afr pending xattrs) to decide whether we can start heal or not. Change-Id: Ibbd56c9aed6026de6ec42422e60293702aaf55f9 BUG: 1408395 Signed-off-by: Ravishankar N Reviewed-on: http://review.gluster.org/16277 NetBSD-regression: NetBSD Build System CentOS-regression: Gluster Build System Smoke: Gluster Build System Reviewed-by: Pranith Kumar Karampuri Tested-by: Pranith Kumar Karampuri --- xlators/cluster/afr/src/afr-common.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/xlators/cluster/afr/src/afr-common.c b/xlators/cluster/afr/src/afr-common.c index e151c01eb0d..46396cf1163 100644 --- a/xlators/cluster/afr/src/afr-common.c +++ b/xlators/cluster/afr/src/afr-common.c @@ -1016,7 +1016,7 @@ afr_replies_interpret (call_frame_t *frame, xlator_t *this, inode_t *inode, for (i = 0; i < priv->child_count; i++) { if (start_heal && priv->child_up[i] && - (!data_readable[i] || !metadata_readable[i])) { + (data_accused[i] || metadata_accused[i])) { *start_heal = _gf_true; break; } -- cgit