summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorshishir gowda <sgowda@redhat.com>2012-12-12 17:54:57 +0530
committerVijay Bellur <vbellur@redhat.com>2012-12-17 12:46:31 -0500
commit8d9a6685068edaf93d08af91932f14072afc2978 (patch)
tree580418b12394ef9eba68d9d06a0fa24a7951ebf6
parent6cd78bb644a2ee1fb34185b5a86002ef4c536173 (diff)
cluster/dht: Add "afr.readdir-failover=off" option the rebalance process
By failing over readdir (default behaviour), rebalance could get duplicate files, as readdir would re-read from offset 0. Rebalance should not attempt to migrate these files again. Additionally, we need to handle these cases as failure in rebalance crawl. BUG: 859387 Change-Id: I77c5c14176bb4d9e593efd6d4739fbc8233bd0c5 Signed-off-by: shishir gowda <sgowda@redhat.com> Reviewed-on: https://code.engineering.redhat.com/gerrit/1991 Reviewed-by: Vijay Bellur <vbellur@redhat.com> Tested-by: Vijay Bellur <vbellur@redhat.com>
-rw-r--r--xlators/cluster/dht/src/dht-rebalance.c34
-rw-r--r--xlators/mgmt/glusterd/src/glusterd-rebalance.c1
2 files changed, 28 insertions, 7 deletions
diff --git a/xlators/cluster/dht/src/dht-rebalance.c b/xlators/cluster/dht/src/dht-rebalance.c
index 30febce..7ac80e7 100644
--- a/xlators/cluster/dht/src/dht-rebalance.c
+++ b/xlators/cluster/dht/src/dht-rebalance.c
@@ -1078,17 +1078,24 @@ gf_defrag_migrate_data (xlator_t *this, gf_defrag_info_t *defrag, loc_t *loc,
while ((ret = syncop_readdirp (this, fd, 131072, offset, NULL,
&entries)) != 0) {
- if (ret < 0)
- break;
+
+ if (ret < 0) {
+
+ gf_log (this->name, GF_LOG_ERROR, "Readdir returned %s."
+ " Aborting migrate-data",
+ strerror(readdir_operrno));
+ goto out;
+ }
/* Need to keep track of ENOENT errno, that means, there is no
need to send more readdirp() */
readdir_operrno = errno;
- free_entries = _gf_true;
-
if (list_empty (&entries.list))
break;
+
+ free_entries = _gf_true;
+
list_for_each_entry_safe (entry, tmp, &entries.list, list) {
if (defrag->defrag_status != GF_DEFRAG_STATUS_STARTED) {
ret = 1;
@@ -1265,6 +1272,7 @@ gf_defrag_fix_layout (xlator_t *this, gf_defrag_info_t *defrag, loc_t *loc,
dict_t *dict = NULL;
off_t offset = 0;
struct iatt iatt = {0,};
+ int readdirp_errno = 0;
ret = syncop_lookup (this, loc, NULL, &iatt, NULL, NULL);
if (ret) {
@@ -1300,12 +1308,22 @@ gf_defrag_fix_layout (xlator_t *this, gf_defrag_info_t *defrag, loc_t *loc,
while ((ret = syncop_readdirp (this, fd, 131072, offset, NULL,
&entries)) != 0)
{
- if ((ret < 0) || (ret && (errno == ENOENT)))
- break;
- free_entries = _gf_true;
+
+ if (ret < 0) {
+ gf_log (this->name, GF_LOG_ERROR, "Readdir returned %s"
+ ". Aborting fix-layout",strerror(errno));
+ goto out;
+ }
+
+ /* Need to keep track of ENOENT errno, that means, there is no
+ need to send more readdirp() */
+ readdirp_errno = errno;
if (list_empty (&entries.list))
break;
+
+ free_entries = _gf_true;
+
list_for_each_entry_safe (entry, tmp, &entries.list, list) {
if (defrag->defrag_status != GF_DEFRAG_STATUS_STARTED) {
ret = 1;
@@ -1381,6 +1399,8 @@ gf_defrag_fix_layout (xlator_t *this, gf_defrag_info_t *defrag, loc_t *loc,
gf_dirent_free (&entries);
free_entries = _gf_false;
INIT_LIST_HEAD (&entries.list);
+ if (readdirp_errno == ENOENT)
+ break;
}
ret = 0;
diff --git a/xlators/mgmt/glusterd/src/glusterd-rebalance.c b/xlators/mgmt/glusterd/src/glusterd-rebalance.c
index b63f464..eac3255 100644
--- a/xlators/mgmt/glusterd/src/glusterd-rebalance.c
+++ b/xlators/mgmt/glusterd/src/glusterd-rebalance.c
@@ -304,6 +304,7 @@ glusterd_handle_defrag_start (glusterd_volinfo_t *volinfo, char *op_errstr,
"--xlator-option",
"*replicate*.metadata-self-heal=off",
"--xlator-option", "*replicate*.entry-self-heal=off",
+ "--xlator-option", "*replicate*.readdir-failover=off",
NULL);
runner_add_arg (&runner, "--xlator-option");
runner_argprintf ( &runner, "*dht.rebalance-cmd=%d",cmd);