Merge pull request #3 from nhs-riak/develop-3.1

Update from martinsumner/leveled
OpenRiak · Nov 8, 2023 · f617dd8 · f617dd8
2 parents 908504f + e2839ba
commit f617dd8
Show file tree

Hide file tree

Showing 20 changed files with 910 additions and 442 deletions.
diff --git a/include/leveled.hrl b/include/leveled.hrl
@@ -1,34 +1,79 @@
-% File paths
+%%%============================================================================
+%%% File paths
+%%%============================================================================
 -define(JOURNAL_FP, "journal").
 -define(LEDGER_FP, "ledger").
+%%%============================================================================
+
+%%%============================================================================
+%%% Configurable startup defaults
+%%%============================================================================
+-define(CACHE_SIZE, 2500).
+-define(MAX_CACHE_MULTTIPLE, 2).
+-define(MIN_CACHE_SIZE, 100).
+-define(MIN_PCL_CACHE_SIZE, 400).
+-define(MAX_PCL_CACHE_SIZE, 28000). 
+    % This is less than actual max - but COIN_SIDECOUNT
+-define(ABSOLUTEMAX_JOURNALSIZE, 4000000000).
+-define(COMPRESSION_METHOD, lz4).
+-define(COMPRESSION_POINT, on_receipt).
+-define(COMPRESSION_LEVEL, 1).
+-define(LOG_LEVEL, info).
+-define(DEFAULT_DBID, 65536).
+-define(OPEN_LASTMOD_RANGE, {0, infinity}).
+-define(SNAPTIMEOUT_SHORT, 900). % 15 minutes
+-define(SNAPTIMEOUT_LONG, 43200). % 12 hours
+-define(SST_PAGECACHELEVEL_NOLOOKUP, 1).
+-define(SST_PAGECACHELEVEL_LOOKUP, 4).
+-define(DEFAULT_STATS_PERC, 10).
+-define(DEFAULT_SYNC_STRATEGY, none).
+%%%============================================================================
+
+%%%============================================================================
+%%% Non-configurable startup defaults
+%%%============================================================================
+-define(MAX_SSTSLOTS, 256).
+-define(LOADING_PAUSE, 1000).
 -define(LOADING_BATCH, 1000).
+-define(CACHE_SIZE_JITTER, 25).
+-define(JOURNAL_SIZE_JITTER, 20).
+-define(LONG_RUNNING, 1000000).
+    % An individual task taking > 1s gets a specific log
+-define(MAX_KEYCHECK_FREQUENCY, 100).
+-define(MIN_KEYCHECK_FREQUENCY, 1).
+-define(MAX_LEVELS, 8).
+    %% Should equal the length of the LEVEL_SCALEFACTOR
+-define(CACHE_TYPE, skpl).
+%%%============================================================================
 
-%% Tag to be used on standard Riak KV objects
+%%%============================================================================
+%%% Tags
+%%%============================================================================
 -define(RIAK_TAG, o_rkv).
-%% Tag to be used on K/V objects for non-Riak purposes
+    %% Tag to be used on standard Riak KV objects
 -define(STD_TAG, o).
-%% Tag used for secondary index keys
+    %% Tag to be used on K/V objects for non-Riak purposes
 -define(IDX_TAG, i).
-%% Tag used for head-only objects
+    %% Tag used for secondary index keys
 -define(HEAD_TAG, h).
+    %% Tag used for head-only objects
 
-%% Inker key type used for 'normal' objects
--define(INKT_STND, stnd). 
-%% Inker key type used for 'batch' objects
+-define(INKT_STND, stnd).
+    %% Inker key type used for 'normal' objects
 -define(INKT_MPUT, mput).
-%% Inker key type used for objects which contain no value, only key changes
-%% This is used currently for objects formed under a 'retain' strategy on Inker
-%% compaction
+    %% Inker key type used for 'batch' objects
 -define(INKT_KEYD, keyd). 
-%% Inker key type used for tombstones
+    %% Inker key type used for objects which contain no value, only key changes
+    %% This is used currently for objects formed under a 'retain' strategy
+    %% on Inker compaction
 -define(INKT_TOMB, tomb).
-
--define(CACHE_TYPE, skpl).
-
--define(MAX_LEVELS, 8).
-%% Should equal the length of the LEVEL_SCALEFACTOR
+    %% Inker key type used for tombstones
+%%%============================================================================
 
 
+%%%============================================================================
+%%% Shared records
+%%%============================================================================
 -record(level,
                         {level :: integer(),
                         is_basement = false :: boolean(),
@@ -47,19 +92,26 @@
                         file_path :: string() | undefined,
                         waste_path :: string() | undefined,
                         binary_mode = false :: boolean(),
-                        sync_strategy = sync,
+                            % Default set by bookie to be `true`
+                            % `false` set here due to legacy of unit tests
+                            % using non-binary keys
+                        sync_strategy = ?DEFAULT_SYNC_STRATEGY,
                         log_options = leveled_log:get_opts() 
                             :: leveled_log:log_options(),
-                        monitor = {no_monitor, 0} :: leveled_monitor:monitor()}).
+                        monitor = {no_monitor, 0}
+                            :: leveled_monitor:monitor()}).
 
 -record(sst_options,
-                        {press_method = native
+                        {press_method = ?COMPRESSION_METHOD
                             :: leveled_sst:press_method(),
+                        press_level = ?COMPRESSION_LEVEL :: non_neg_integer(),
                         log_options = leveled_log:get_opts() 
                             :: leveled_log:log_options(),
-                        max_sstslots = 256 :: pos_integer(),
-                        pagecache_level = 1 :: pos_integer(),
-                        monitor = {no_monitor, 0} :: leveled_monitor:monitor()}).
+                        max_sstslots = ?MAX_SSTSLOTS :: pos_integer(),
+                        pagecache_level = ?SST_PAGECACHELEVEL_NOLOOKUP
+                            :: pos_integer(),
+                        monitor = {no_monitor, 0}
+                            :: leveled_monitor:monitor()}).
 
 -record(inker_options,
                         {cdb_max_size :: integer() | undefined,
@@ -70,14 +122,16 @@
                         source_inker :: pid() | undefined,
                         reload_strategy = [] :: list(),
                         waste_retention_period :: integer() | undefined,
-                        compression_method = native :: lz4|native,
+                        compression_method = ?COMPRESSION_METHOD
+                            :: lz4|native|none,
                         compress_on_receipt = false :: boolean(),
                         max_run_length,
                         singlefile_compactionperc :: float()|undefined,
                         maxrunlength_compactionperc :: float()|undefined,
                         score_onein = 1 :: pos_integer(),
                         snaptimeout_long :: pos_integer() | undefined,
-                        monitor = {no_monitor, 0} :: leveled_monitor:monitor()}).
+                        monitor = {no_monitor, 0}
+                            :: leveled_monitor:monitor()}).
 
 -record(penciller_options,
                         {root_path :: string() | undefined,
@@ -89,19 +143,23 @@
                         bookies_mem :: tuple() | undefined,
                         source_penciller :: pid() | undefined,
                         snapshot_longrunning = true :: boolean(),
-                        compression_method = native :: lz4|native,
+                        compression_method = ?COMPRESSION_METHOD
+                            :: lz4|native|none,
                         levelzero_cointoss = false :: boolean(),
                         snaptimeout_short :: pos_integer() | undefined,
                         snaptimeout_long :: pos_integer() | undefined,
-                        monitor = {no_monitor, 0} :: leveled_monitor:monitor()}).
+                        monitor = {no_monitor, 0}
+                            :: leveled_monitor:monitor()}).
 
 -record(iclerk_options,
                         {inker :: pid() | undefined,
                          max_run_length :: integer() | undefined,
                          cdb_options = #cdb_options{} :: #cdb_options{},
                          waste_retention_period :: integer() | undefined,
-                         compression_method = native :: lz4|native,
+                         compression_method = ?COMPRESSION_METHOD
+                            :: lz4|native|none,
                          singlefile_compactionperc :: float()|undefined,
                          maxrunlength_compactionperc :: float()|undefined,
                          score_onein = 1 :: pos_integer(),
                          reload_strategy = [] :: list()}).
+%%%============================================================================
diff --git a/priv/leveled.schema b/priv/leveled.schema
@@ -46,11 +46,10 @@
 %% Can be lz4 or native (which will use the Erlang native zlib compression)
 %% within term_to_binary
 {mapping, "leveled.compression_method", "leveled.compression_method", [
-  {default, native},
-  {datatype, atom}
+  {datatype, {enum, [native, lz4, none]}},
+  {default, native}
 ]}.
 
-
 %% @doc Compression point
 %% The point at which compression is applied to the Journal (the Ledger is 
 %% always compressed).  Use on_receipt or on_compact.  on_compact is suitable
@@ -61,6 +60,16 @@
   {datatype, atom}
 ]}.
 
+%% @doc Compresstion Level (Ledger LSM)
+%% Specify the level in the LSM tree from which compression should occur.
+%% Defaults to L1, so only L0 writes are uncompressed.
+{mapping, "leveled.compression_level", "leveled.compression_level", [
+  {default, 1},
+  {datatype, integer},
+  {validators, ["range:0-7"]},
+  hidden
+]}.
+
 %% @doc Log level
 %% Can be debug, info, warn, error or critical
 %% Set the minimum log level to be used within leveled.  Leveled will log many

diff --git a/priv/leveled_multi.schema b/priv/leveled_multi.schema
@@ -39,8 +39,8 @@
 %% Can be lz4 or native (which will use the Erlang native zlib compression)
 %% within term_to_binary
 {mapping, "multi_backend.$name.leveled.compression_method", "riak_kv.multi_backend", [
-  {default, lz4},
-  {datatype, atom},
+  {datatype, {enum, [native, lz4, none]}},
+  {default, native}
   hidden
 ]}.
 
@@ -55,6 +55,16 @@
   hidden
 ]}.
 
+%% @doc Compresstion Level (Ledger LSM)
+%% Specify the level in the LSM tree from which compression should occur.
+%% Defaults to L1, so only L0 writes are uncompressed.
+{mapping, "multi_backend.$name.leveled.compression_level", "riak_kv.multi_backend", [
+  {default, 1},
+  {datatype, integer},
+  {validators, ["range:0-7"]},
+  hidden
+]}.
+
 
 %% @doc The approximate size (in bytes) when a Journal file should be rolled.
 %% Normally keep this as around the size of o(100K) objects.  Default is 1GB.