@RunLLM get earliest in backlog

To get the earliest time in backlog for a Pulsar topic, use the

pulsar-admin topics stats

command with the

--get-earliest-time-in-backlog

(or

-etb

) option. This will return the publish time of the oldest unacknowledged message in the backlog. For example:

pulsar-admin topics stats my-topic --get-earliest-time-in-backlog

This option is also available via the REST API and Java admin client. For partitioned topics, use

partitioned-stats

with the same option. More details are in the official docs and code references: Pulsar Admin API Topics, CmdTopics.java, and GetStatsOptions.java. Would you like more detail or code examples for a specific client? Sources: - Pulsar Admin API Topics - CmdTopics.java - GetStatsOptions.java 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.

@RunLLM when auto recovery is not configured will replication worker come in book

If auto recovery is not configured or enabled, the ReplicationWorker will not run or participate in the BookKeeper cluster. ReplicationWorker is a component of AutoRecovery, and both are only active when AutoRecovery is enabled in the configuration or started as a separate process. Without AutoRecovery, ledgers on failed bookies will not be automatically rereplicated by ReplicationWorker or any other process. Would you like more detail on how to enable or configure AutoRecovery? Sources: - AutoRecovery architecture - Using AutoRecovery 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.

@RunLLM 1757658138.239131742,"2025-09-11T232218.239131742-07:00","2025-09-12T062218,238 [ReplicationWorker] ERROR org.apache.bookkeeper.replication.ReplicationWorker - ReplicationWorker failed to replicate Ledger : 549343 for 158 number of times, so deferring the ledger lock release by 300000 msecs","itomdipulsar-bookkeeper-0","op","itomdipulsar-bookkeeper"," 1757658138.239178642,"2025-09-11T232218.239178642-07:00","2025-09-12T062218,238 [ReplicationWorker] WARN org.apache.bookkeeper.replication.ReplicationWorker - failed while replicating fragments","itomdipulsar-bookkeeper-0","","itomdipulsar-bookkeeper"

The log lines indicate that

org.apache.bookkeeper.replication.ReplicationWorker.run

is repeatedly failing to replicate ledger 549343, and after 158 failed attempts, it is deferring the ledger lock release by 300,000 milliseconds (5 minutes). This is standard behavior: after each failure, ReplicationWorker increases the delay before retrying, up to a configured maximum, to avoid overwhelming the system. The warning "failed while replicating fragments" confirms the replication attempt failed in the main run loop. Would you like more detail on why this happens or how to troubleshoot it? Sources: - ReplicationWorker.java (run method, error handling, and backoff) 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.

@RunLLM replication worker in bookie or auto recovery

ReplicationWorker is part of the AutoRecovery process, not the core Bookie process. It runs as a thread within AutoRecoveryMain, which can be started either on the same machine as a Bookie or as a separate service. If AutoRecovery is not enabled, ReplicationWorker does not run. Would you like more detail on how to configure or deploy AutoRecovery? Sources: - AutoRecovery architecture - Using AutoRecovery 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.

@RunLLM replication worker errors in bookkeeper

ReplicationWorker errors in BookKeeper typically occur when it cannot rereplicate under-replicated ledgers, often due to not enough available bookies, read/write failures, or data loss. Common errors include

BKNotEnoughBookiesException

, repeated failed replication attempts, and warnings like "failed while replicating fragments" in logs. These errors trigger exponential backoff and deferred lock release to avoid system overload. Would you like more detail on specific error types or troubleshooting steps? Sources: - ReplicationWorker.java (error handling) - Bookkeeper issue: ReplicationWorker not work problem - Bookkeeper issue: Autorecovery stuck when ensemble equals total bookies 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.