cluster/tier: Adding compaction option for metadata databases

Problem: As metadata in the database fills up, querying the database
take a long time. As a result, tier migration slows down.  To
counteract this, we added a way to enable the compaction methods of
the underlying database. The goal is to reduce the size of the
underlying file by eliminating database fragmentation.

NOTE: There is currently a bug where sometimes a brick will
attempt to activate compaction. This happens even compaction is already
turned on.

The cause is narrowed down to the compact_mode_switch flipping its value.

Changes: libglusterfs/src/gfdb - Added a gfdb function to compact the
underlying database, compact_db() This is a no-op if the database has
no such option.

- Added a compaction function for SQLite3 that does the following

1) Changes the auto_vacuum pragma of the database
2) Compacts the database according to the type of compaction requested

- Compaction type can be changed by changing the macro
  GF_SQL_COMPACT_DEF to one of the 4 compaction types in
  gfdb_sqlite3.h

  It is currently set to GF_SQL_COMPACT_INCR, or incremental
  vacuuming.

xlators/cluster/dht/src - Added the following command-line option to
enable SQLite3 compaction.

gluster volume set <vol-name> tier-compact <off|on>

- Added the following command-line option to change the frequency the
  hot and cold tier are ordered to compact.

gluster volume set <vol-name> tier-hot-compact-frequency <int>
gluster volume set <vol-name> tier-cold-compact-frequency <int>

- tier daemon periodically sends the (new)
  GFDB_IPC_CTR_SET_COMPACT_PRAGMA IPC to the CTR xlator. The IPC
  triggers compaction of the database.

  The inputs are both gf_boolean_t.

  IPC Input:

  compact_active: Is compaction currently on for the db.
  compact_mode_switched: Did we flip the compaction switch recently?

  IPC Output:

  0 if the compaction succeeds.
  Non-zero otherwise.

xlators/features/changetimerecorder/src/ - When the CTR gets the
compaction IPC, it launches a thread that will perform the
compaction. The IPC ends after the thread is launched. To avoid extra
allocations, the parameters are passed using static variables.

Change-Id: I5e1433becb9eeff2afe8dcb4a5798977bf5ba0dd
Signed-off-by: Diogenes Nunez <dnunez@redhat.com>
Reviewed-on: http://review.gluster.org/15031
Reviewed-by: Milind Changire <mchangir@redhat.com>
Reviewed-by: Dan Lambright <dlambrig@redhat.com>
Tested-by: Dan Lambright <dlambrig@redhat.com>
CentOS-regression: Gluster Build System <jenkins@build.gluster.org>
NetBSD-regression: NetBSD Build System <jenkins@build.gluster.org>
Smoke: Gluster Build System <jenkins@build.gluster.org>

1 files changed, 20 insertions, 6 deletions

diff --git a/xlators/cluster/dht/src/dht-shared.c b/xlators/cluster/dht/src/dht-shared.c
index 873ced53eec..f410f71b5a6 100644
--- a/xlators/cluster/dht/src/dht-shared.c
+++ b/xlators/cluster/dht/src/dht-shared.c
@@ -148,7 +148,7 @@ dht_priv_dump (xlator_t *this)
         gf_proc_dump_write("search_unhashed", "%d", conf->search_unhashed);
         gf_proc_dump_write("gen", "%d", conf->gen);
         gf_proc_dump_write("min_free_disk", "%lf", conf->min_free_disk);
-	gf_proc_dump_write("min_free_inodes", "%lf", conf->min_free_inodes);
+        gf_proc_dump_write("min_free_inodes", "%lf", conf->min_free_inodes);
         gf_proc_dump_write("disk_unit", "%c", conf->disk_unit);
         gf_proc_dump_write("refresh_interval", "%d", conf->refresh_interval);
         gf_proc_dump_write("unhashed_sticky_bit", "%d", conf->unhashed_sticky_bit);
@@ -433,14 +433,14 @@ dht_reconfigure (xlator_t *this, dict_t *options)
         GF_OPTION_RECONF ("lookup-optimize", conf->lookup_optimize, options,
                           bool, out);
 
-	GF_OPTION_RECONF ("min-free-disk", conf->min_free_disk, options,
+        GF_OPTION_RECONF ("min-free-disk", conf->min_free_disk, options,
                           percent_or_size, out);
         /* option can be any one of percent or bytes */
         conf->disk_unit = 0;
         if (conf->min_free_disk < 100.0)
                 conf->disk_unit = 'p';
 
-	GF_OPTION_RECONF ("min-free-inodes", conf->min_free_inodes, options,
+        GF_OPTION_RECONF ("min-free-inodes", conf->min_free_inodes, options,
                           percent, out);
 
         GF_OPTION_RECONF ("directory-layout-spread", conf->dir_spread_cnt,
@@ -711,8 +711,8 @@ dht_init (xlator_t *this)
 
         GF_OPTION_INIT ("use-readdirp", conf->use_readdirp, bool, err);
 
-	GF_OPTION_INIT ("min-free-disk", conf->min_free_disk, percent_or_size,
-			err);
+        GF_OPTION_INIT ("min-free-disk", conf->min_free_disk, percent_or_size,
+                        err);
 
         GF_OPTION_INIT ("min-free-inodes", conf->min_free_inodes, percent,
                         err);
@@ -901,7 +901,7 @@ struct volume_options options[] = {
           "process starts balancing out the cluster, and logs will appear "
           "in log files",
         },
-	{ .key  = {"min-free-inodes"},
+        { .key  = {"min-free-inodes"},
           .type = GF_OPTION_TYPE_PERCENT,
           .default_value = "5%",
           .description = "after system has only N% of inodes, warnings "
@@ -1038,6 +1038,20 @@ struct volume_options options[] = {
           .type = GF_OPTION_TYPE_STR,
           .default_value = "test",
         },
+        { .key         = {"tier-compact"},
+          .type = GF_OPTION_TYPE_BOOL,
+          .default_value = "off",
+        },
+        { .key         = {"tier-hot-compact-frequency"},
+          .type = GF_OPTION_TYPE_INT,
+          .default_value = "604800",
+          .description = "Frequency to compact DBs on hot tier in system"
+        },
+        { .key         = {"tier-cold-compact-frequency"},
+          .type = GF_OPTION_TYPE_INT,
+          .default_value = "604800",
+          .description = "Frequency to compact DBs on cold tier in system"
+        },
         { .key         = {"tier-max-mb"},
           .type = GF_OPTION_TYPE_INT,
           .default_value = "4000",