cluster/tier: add tunable to migrate files based on size

This fix adds a paramater "tier-max_promote_size" to control wether
a file is migrated or not based on its size. By default the value
is 0, meaning all files are migrated. If set to a non-zero
value, files larger than the parameter won't be moved
in tiered volumes.

Change-Id: Ia6b88e9b2508935bef500d956f9192e59670fe00
BUG: 1313495
Signed-off-by: Dan Lambright <dlambrig@redhat.com>
Reviewed-on: http://review.gluster.org/13570
Smoke: Gluster Build System <jenkins@build.gluster.com>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
CentOS-regression: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Joseph Fernandes

5 files changed, 41 insertions, 0 deletions

diff --git a/tests/basic/tier/tier.t b/tests/basic/tier/tier.t
index 5874a9b6e83..deff2e299ef 100755
--- a/tests/basic/tier/tier.t
+++ b/tests/basic/tier/tier.t
@@ -112,6 +112,7 @@ TEST $CLI volume set $V0 cluster.watermark-hi 85
 TEST $CLI volume set $V0 cluster.watermark-low 75
 TEST $CLI volume set $V0 cluster.tier-max-mb 1000
 TEST $CLI volume set $V0 cluster.tier-max-files 1000
+TEST $CLI volume set $V0 cluster.tier-max-promote-file-size 1000
 TEST ! $CLI volume set $V0 cluster.tier-max-files -3
 TEST ! $CLI volume set $V0 cluster.watermark-low 90
 TEST ! $CLI volume set $V0 cluster.read-freq-threshold -12
diff --git a/xlators/cluster/dht/src/dht-common.h b/xlators/cluster/dht/src/dht-common.h
index d9b52e28bb9..6ba94ec8814 100644
--- a/xlators/cluster/dht/src/dht-common.h
+++ b/xlators/cluster/dht/src/dht-common.h
@@ -357,6 +357,7 @@ typedef struct gf_tier_conf {
         uint64_t                     max_migrate_bytes;
         int                          max_migrate_files;
         tier_mode_t                  mode;
+        int                          tier_max_promote_size;
         int                          tier_promote_frequency;
         int                          tier_demote_frequency;
         uint64_t                     st_last_promoted_size;
diff --git a/xlators/cluster/dht/src/dht-shared.c b/xlators/cluster/dht/src/dht-shared.c
index e3e7fadca3f..856819ac052 100644
--- a/xlators/cluster/dht/src/dht-shared.c
+++ b/xlators/cluster/dht/src/dht-shared.c
@@ -1032,6 +1032,10 @@ struct volume_options options[] = {
           .type = GF_OPTION_TYPE_INT,
           .default_value = "4000",
         },
+        { .key         = {"tier-max-promote-file-size"},
+          .type = GF_OPTION_TYPE_INT,
+          .default_value = "0",
+        },
         { .key         = {"tier-max-files"},
           .type = GF_OPTION_TYPE_INT,
           .default_value = "10000",
diff --git a/xlators/cluster/dht/src/tier.c b/xlators/cluster/dht/src/tier.c
index 007cfc60c5a..b99497bc66a 100644
--- a/xlators/cluster/dht/src/tier.c
+++ b/xlators/cluster/dht/src/tier.c
@@ -724,6 +724,17 @@ tier_migrate_using_query_file (void *_args)
                                 per_link_status = -1;
                                 goto abort;
                         }
+
+                        if (query_cbk_args->is_promotion &&
+                            defrag->tier_conf.tier_max_promote_size &&
+                            (current.ia_size > defrag->tier_conf.tier_max_promote_size)) {
+                                gf_msg (this->name, GF_LOG_INFO, 0,
+                                        DHT_MSG_LOG_TIER_STATUS,
+                                        "File size exceeds maxsize for promotion. ");
+                                per_link_status = 1;
+                                goto abort;
+                        }
+
                         linked_inode = inode_link (loc.inode, NULL, NULL,
                                                         &current);
                         inode_unref (loc.inode);
@@ -2079,6 +2090,7 @@ tier_init (xlator_t *this)
 {
         int               ret            = -1;
         int               freq           = 0;
+        int               maxsize        = 0;
         dht_conf_t       *conf           = NULL;
         gf_defrag_info_t *defrag         = NULL;
         char             *voldir         = NULL;
@@ -2130,6 +2142,14 @@ tier_init (xlator_t *this)
         defrag->tier_conf.is_tier = 1;
 
         ret = dict_get_int32 (this->options,
+                              "tier-max-promote-file-size", &maxsize);
+        if (ret) {
+                maxsize = 0;
+        }
+
+        defrag->tier_conf.tier_max_promote_size = maxsize;
+
+        ret = dict_get_int32 (this->options,
                               "tier-promote-frequency", &freq);
         if (ret) {
                 freq = DEFAULT_PROMOTE_FREQ_SEC;
@@ -2316,6 +2336,10 @@ tier_reconfigure (xlator_t *this, dict_t *options)
 
         if (conf->defrag) {
                 defrag = conf->defrag;
+                GF_OPTION_RECONF ("tier-max-promote-file-size",
+                                  defrag->tier_conf.tier_max_promote_size,
+                                  options, int32, out);
+
                 GF_OPTION_RECONF ("tier-promote-frequency",
                                   defrag->tier_conf.tier_promote_frequency,
                                   options, int32, out);
diff --git a/xlators/mgmt/glusterd/src/glusterd-volume-set.c b/xlators/mgmt/glusterd/src/glusterd-volume-set.c
index 5d9aafeb661..d0e5efbd9a0 100644
--- a/xlators/mgmt/glusterd/src/glusterd-volume-set.c
+++ b/xlators/mgmt/glusterd/src/glusterd-volume-set.c
@@ -452,6 +452,7 @@ validate_tier (glusterd_volinfo_t *volinfo, dict_t *dict, char *key,
                 }
         } else if (strstr (key, "tier-promote-frequency") ||
                    strstr (key, "tier-max-mb") ||
+                   strstr (key, "tier-max-promote-file-size") ||
                    strstr (key, "tier-max-files") ||
                    strstr (key, "tier-demote-frequency")) {
                 if (origin_val < 1) {
@@ -2496,6 +2497,16 @@ struct volopt_map_entry glusterd_volopt_map[] = {
           " Cache mode does so based on whether the cache is full or not,"
           " as specified with watermarks."
         },
+        { .key         = "cluster.tier-max-promote-file-size",
+          .voltype     = "cluster/tier",
+          .option      = "tier-max-promote-file-size",
+          .value       = "0",
+          .op_version  = GD_OP_VERSION_3_7_10,
+          .flags       = OPT_FLAG_CLIENT_OPT,
+          .validate_fn = validate_tier,
+          .description = "The maximum file size in bytes that is promoted. If 0, there"
+          " is no maximum size (default)."
+        },
         { .key         = "cluster.tier-max-mb",
           .voltype     = "cluster/tier",
           .option      = "tier-max-mb",