This is an automated email from the ASF dual-hosted git repository.
github-actions[bot] pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/pulsar-site.git
The following commit(s) were added to refs/heads/main by this push:
new 5d5fc5b71b0 Docs sync done from apache/pulsar (#982eac3)
5d5fc5b71b0 is described below
commit 5d5fc5b71b0a082ce6d56d7e26208b69e8de4c79
Author: github-actions[bot]
<41898282+github-actions[bot]@users.noreply.github.com>
AuthorDate: Thu Jun 11 01:53:38 2026 +0000
Docs sync done from apache/pulsar (#982eac3)
---
.../next/config/reference-configuration-broker.md | 20 +++++++++++---------
.../config/reference-configuration-standalone.md | 20 +++++++++++---------
.../next/pulsar/initialize-cluster-metadata.md | 2 +-
3 files changed, 23 insertions(+), 19 deletions(-)
diff --git a/static/reference/next/config/reference-configuration-broker.md
b/static/reference/next/config/reference-configuration-broker.md
index 2cc82520aaf..49a6a77f6aa 100644
--- a/static/reference/next/config/reference-configuration-broker.md
+++ b/static/reference/next/config/reference-configuration-broker.md
@@ -2132,7 +2132,7 @@ Setting this value to 0 will disable the limit calculated
per consumer.
**Type**: `int`
-**Default**: `2000`
+**Default**: `4000`
**Dynamic**: `true`
@@ -2141,13 +2141,14 @@ Setting this value to 0 will disable the limit
calculated per consumer.
### keySharedLookAheadMsgInReplayThresholdPerSubscription
For Key_Shared subscriptions, if messages cannot be dispatched to consumers
due to a slow consumer or a blocked key hash (because of ordering constraints),
the broker will continue reading more messages from the backlog and attempt to
dispatch them to consumers until the number of replay messages reaches the
calculated threshold.
Formula: threshold = min(keySharedLookAheadMsgInReplayThresholdPerConsumer *
connected consumer count,
keySharedLookAheadMsgInReplayThresholdPerSubscription).
-This value should be set to a value less than 2 *
managedLedgerMaxUnackedRangesToPersist.
-Setting this value to 0 will disable the limit calculated per subscription.
+Keep this value relatively low to increase the cache hit ratio for Key_Shared
replay queue reads; it should also be less than 2 *
managedLedgerMaxUnackedRangesToPersist.
+However, with workloads that have low key cardinality (few distinct keys), a
low value can leave some consumers idle, since the dispatcher pauses once this
threshold is reached and stops reading new messages for the other consumers;
increasing the value allows more messages to be in flight in such cases. Since
the effective threshold is the minimum of the per-consumer and per-subscription
limits, keySharedLookAheadMsgInReplayThresholdPerConsumer should also be
increased, or set to 0 (dis [...]
+Setting this value to 0 will disable the limit calculated per subscription.
Disabling it might result in the number of unacked ranges to persist exceeding
managedLedgerMaxUnackedRangesToPersist, therefore disabling is not recommended.
**Type**: `int`
-**Default**: `20000`
+**Default**: `40000`
**Dynamic**: `true`
@@ -5273,7 +5274,7 @@ If value is NONE, then save the ManagedCursorInfo bytes
data directly.
**Type**: `java.lang.String`
-**Default**: `NONE`
+**Default**: `LZ4`
**Dynamic**: `false`
@@ -5606,7 +5607,7 @@ If value is invalid or NONE, then save the
ManagedLedgerInfo bytes data directly
**Type**: `java.lang.String`
-**Default**: `NONE`
+**Default**: `LZ4`
**Dynamic**: `false`
@@ -5642,7 +5643,7 @@ When this limit is exceeded, remaining batch message
containing the batch delete
**Type**: `int`
-**Default**: `10000`
+**Default**: `200000`
**Dynamic**: `false`
@@ -5734,10 +5735,11 @@ Maximum ledger size before triggering a rollover for a
topic (MB)
Max number of `acknowledgment holes` that are going to be persistently stored.
When acknowledging out of order, a consumer will leave holes that are supposed
to be quickly filled by acking all the messages. The information of which
messages are acknowledged is persisted by compressing in `ranges` of messages
that were acknowledged. After the max number of ranges is reached, the
information will only be tracked in memory and messages will be redelivered in
case of crashes.
+Note: when managedLedgerPersistIndividualAckAsLongArray is enabled (the
default), the persisted size is bounded by the backlog size (the range of
entries the cursor spans), not by this max number of unacked ranges. For
BookKeeper ledger storage, with the default broker maxMessageSize and
BookKeeper nettyMaxFrameSizeBytes, the state fits a backlog of about 30M
entries (excluding the managedLedgerMaxBatchDeletedIndexToPersist storage,
whose size is instead relative to the number of acknowl [...]
**Type**: `int`
-**Default**: `10000`
+**Default**: `200000`
**Dynamic**: `false`
@@ -5750,7 +5752,7 @@ If number of unack message range is higher than this
limit then broker will pers
**Type**: `int`
-**Default**: `1000`
+**Default**: `200000`
**Dynamic**: `false`
diff --git a/static/reference/next/config/reference-configuration-standalone.md
b/static/reference/next/config/reference-configuration-standalone.md
index 2cc82520aaf..49a6a77f6aa 100644
--- a/static/reference/next/config/reference-configuration-standalone.md
+++ b/static/reference/next/config/reference-configuration-standalone.md
@@ -2132,7 +2132,7 @@ Setting this value to 0 will disable the limit calculated
per consumer.
**Type**: `int`
-**Default**: `2000`
+**Default**: `4000`
**Dynamic**: `true`
@@ -2141,13 +2141,14 @@ Setting this value to 0 will disable the limit
calculated per consumer.
### keySharedLookAheadMsgInReplayThresholdPerSubscription
For Key_Shared subscriptions, if messages cannot be dispatched to consumers
due to a slow consumer or a blocked key hash (because of ordering constraints),
the broker will continue reading more messages from the backlog and attempt to
dispatch them to consumers until the number of replay messages reaches the
calculated threshold.
Formula: threshold = min(keySharedLookAheadMsgInReplayThresholdPerConsumer *
connected consumer count,
keySharedLookAheadMsgInReplayThresholdPerSubscription).
-This value should be set to a value less than 2 *
managedLedgerMaxUnackedRangesToPersist.
-Setting this value to 0 will disable the limit calculated per subscription.
+Keep this value relatively low to increase the cache hit ratio for Key_Shared
replay queue reads; it should also be less than 2 *
managedLedgerMaxUnackedRangesToPersist.
+However, with workloads that have low key cardinality (few distinct keys), a
low value can leave some consumers idle, since the dispatcher pauses once this
threshold is reached and stops reading new messages for the other consumers;
increasing the value allows more messages to be in flight in such cases. Since
the effective threshold is the minimum of the per-consumer and per-subscription
limits, keySharedLookAheadMsgInReplayThresholdPerConsumer should also be
increased, or set to 0 (dis [...]
+Setting this value to 0 will disable the limit calculated per subscription.
Disabling it might result in the number of unacked ranges to persist exceeding
managedLedgerMaxUnackedRangesToPersist, therefore disabling is not recommended.
**Type**: `int`
-**Default**: `20000`
+**Default**: `40000`
**Dynamic**: `true`
@@ -5273,7 +5274,7 @@ If value is NONE, then save the ManagedCursorInfo bytes
data directly.
**Type**: `java.lang.String`
-**Default**: `NONE`
+**Default**: `LZ4`
**Dynamic**: `false`
@@ -5606,7 +5607,7 @@ If value is invalid or NONE, then save the
ManagedLedgerInfo bytes data directly
**Type**: `java.lang.String`
-**Default**: `NONE`
+**Default**: `LZ4`
**Dynamic**: `false`
@@ -5642,7 +5643,7 @@ When this limit is exceeded, remaining batch message
containing the batch delete
**Type**: `int`
-**Default**: `10000`
+**Default**: `200000`
**Dynamic**: `false`
@@ -5734,10 +5735,11 @@ Maximum ledger size before triggering a rollover for a
topic (MB)
Max number of `acknowledgment holes` that are going to be persistently stored.
When acknowledging out of order, a consumer will leave holes that are supposed
to be quickly filled by acking all the messages. The information of which
messages are acknowledged is persisted by compressing in `ranges` of messages
that were acknowledged. After the max number of ranges is reached, the
information will only be tracked in memory and messages will be redelivered in
case of crashes.
+Note: when managedLedgerPersistIndividualAckAsLongArray is enabled (the
default), the persisted size is bounded by the backlog size (the range of
entries the cursor spans), not by this max number of unacked ranges. For
BookKeeper ledger storage, with the default broker maxMessageSize and
BookKeeper nettyMaxFrameSizeBytes, the state fits a backlog of about 30M
entries (excluding the managedLedgerMaxBatchDeletedIndexToPersist storage,
whose size is instead relative to the number of acknowl [...]
**Type**: `int`
-**Default**: `10000`
+**Default**: `200000`
**Dynamic**: `false`
@@ -5750,7 +5752,7 @@ If number of unack message range is higher than this
limit then broker will pers
**Type**: `int`
-**Default**: `1000`
+**Default**: `200000`
**Dynamic**: `false`
diff --git a/static/reference/next/pulsar/initialize-cluster-metadata.md
b/static/reference/next/pulsar/initialize-cluster-metadata.md
index 30ea6f2db2c..b7e24d0a6b1 100644
--- a/static/reference/next/pulsar/initialize-cluster-metadata.md
+++ b/static/reference/next/pulsar/initialize-cluster-metadata.md
@@ -1,4 +1,4 @@
-2026-06-10T01:45:31,898+0000 [main] INFO
org.apache.bookkeeper.meta.MetadataDrivers - BookKeeper metadata driver manager
initialized {}
+2026-06-11T01:53:21,893+0000 [main] INFO
org.apache.bookkeeper.meta.MetadataDrivers - BookKeeper metadata driver manager
initialized {}
# initialize-cluster-metadata
