Skip to content

Commit 1cc2402

Browse files
authored
[improve][broker] Add managedCursor/LedgerInfoCompressionType settings to broker.conf (apache#24391)
1 parent d1bca65 commit 1cc2402

File tree

2 files changed

+46
-0
lines changed

2 files changed

+46
-0
lines changed

conf/broker.conf

Lines changed: 23 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1319,6 +1319,29 @@ managedLedgerUnackedRangesOpenCacheSetEnabled=true
13191319
# MetadataStore.
13201320
managedLedgerMaxUnackedRangesToPersistInMetadataStore=1000
13211321

1322+
# ManagedCursorInfo compression type, option values (NONE, LZ4, ZLIB, ZSTD, SNAPPY).
1323+
# If value is NONE, then save the ManagedCursorInfo bytes data directly without compression.
1324+
# Using compression reduces the size of persistent cursor (subscription) metadata. This enables using a higher
1325+
# managedLedgerMaxUnackedRangesToPersistInMetadataStore value and reduces the overall metadata stored in
1326+
# the metadata store such as ZooKeeper.
1327+
managedCursorInfoCompressionType=NONE
1328+
1329+
# ManagedCursorInfo compression size threshold (bytes), only compress metadata when origin size more then this value.
1330+
# 0 means compression will always apply.
1331+
managedCursorInfoCompressionThresholdInBytes=16384
1332+
1333+
# ManagedLedgerInfo compression type, option values (NONE, LZ4, ZLIB, ZSTD, SNAPPY).
1334+
# If value is invalid or NONE, then save the ManagedLedgerInfo bytes data directly without compression.
1335+
# Using compression reduces the size of the persistent topic metadata. When a topic contains a large number of
1336+
# individual ledgers in BookKeeper or tiered storage, compression helps prevent the metadata size from exceeding
1337+
# the maximum size of a metadata store entry (ZNode in ZooKeeper). This also reduces the overall metadata stored
1338+
# in the metadata store such as ZooKeeper.
1339+
managedLedgerInfoCompressionType=NONE
1340+
1341+
# ManagedLedgerInfo compression size threshold (bytes), only compress metadata when origin size more then this value.
1342+
# 0 means compression will always apply.
1343+
managedLedgerInfoCompressionThresholdInBytes=16384
1344+
13221345
# Skip reading non-recoverable/unreadable data-ledger under managed-ledger's list. It helps when data-ledgers gets
13231346
# corrupted at bookkeeper and managed-cursor is stuck at that ledger.
13241347
autoSkipNonRecoverableData=false

conf/standalone.conf

Lines changed: 23 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -849,6 +849,29 @@ managedLedgerUnackedRangesOpenCacheSetEnabled=true
849849
# MetadataStore.
850850
managedLedgerMaxUnackedRangesToPersistInMetadataStore=1000
851851

852+
# ManagedCursorInfo compression type, option values (NONE, LZ4, ZLIB, ZSTD, SNAPPY).
853+
# If value is NONE, then save the ManagedCursorInfo bytes data directly without compression.
854+
# Using compression reduces the size of persistent cursor (subscription) metadata. This enables using a higher
855+
# managedLedgerMaxUnackedRangesToPersistInMetadataStore value and reduces the overall metadata stored in
856+
# the metadata store such as ZooKeeper.
857+
managedCursorInfoCompressionType=NONE
858+
859+
# ManagedCursorInfo compression size threshold (bytes), only compress metadata when origin size more then this value.
860+
# 0 means compression will always apply.
861+
managedCursorInfoCompressionThresholdInBytes=16384
862+
863+
# ManagedLedgerInfo compression type, option values (NONE, LZ4, ZLIB, ZSTD, SNAPPY).
864+
# If value is invalid or NONE, then save the ManagedLedgerInfo bytes data directly without compression.
865+
# Using compression reduces the size of the persistent topic metadata. When a topic contains a large number of
866+
# individual ledgers in BookKeeper or tiered storage, compression helps prevent the metadata size from exceeding
867+
# the maximum size of a metadata store entry (ZNode in ZooKeeper). This also reduces the overall metadata stored
868+
# in the metadata store such as ZooKeeper.
869+
managedLedgerInfoCompressionType=NONE
870+
871+
# ManagedLedgerInfo compression size threshold (bytes), only compress metadata when origin size more then this value.
872+
# 0 means compression will always apply.
873+
managedLedgerInfoCompressionThresholdInBytes=16384
874+
852875
# Skip reading non-recoverable/unreadable data-ledger under managed-ledger's list. It helps when data-ledgers gets
853876
# corrupted at bookkeeper and managed-cursor is stuck at that ledger.
854877
autoSkipNonRecoverableData=false

0 commit comments

Comments
 (0)