[
https://issues.apache.org/jira/browse/AMQ-8617?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Endre Stølsvik updated AMQ-8617:
--------------------------------
Description:
Scenario on client:
# Employing RedeliveryPolicy with exponential backoff (keeping maximum
redeliveries at default 6)
# Enabled non-blocking redelivery
# Receiving e.g. 100 consecutive poison messages which are rolled back
repeatedly (which eventually should DLQ after max redeliveries)
This will result in massive redelivery delays due to a logic bug.
The reason is that redeliveryDelay is a field variable kept on the
ActiveMQMessageConsumer, instead of being a property on the message - or that
the redelivery delay was calculated per message based on the redelivery count.
When consecutive messages rollbacks multiple times, the redeliveryDelay field
is continuously multiplied by the backoff multiplier, resulting in enormous
delays.
Pull Request for fix, including test:
https://github.com/apache/activemq/pull/843
was:
Scenario on client:
# Employing RedeliveryPolicy with exponential backoff (keeping maximum
redeliveries at default 6)
# Enabled non-blocking redelivery
# Receiving e.g. 100 consecutive poison messages which are rolled back
repeatedly (which eventually should DLQ after max redeliveries)
This will result in massive redelivery delays due to a logic bug.
The reason is that redeliveryDelay is a field variable kept on the
ActiveMQMessageConsumer, instead of being a property on the message - or that
the redelivery delay was calculated per message based on the redelivery count.
When consecutive messages rollbacks multiple times, the redeliveryDelay field
is continuously multiplied by the backoff multiplier, resulting in enormous
delays.
I'll submit a PR.
> RedeliveryPolicy:Exponential Backoff + NonBlockingRedelivery = too long delays
> ------------------------------------------------------------------------------
>
> Key: AMQ-8617
> URL: https://issues.apache.org/jira/browse/AMQ-8617
> Project: ActiveMQ
> Issue Type: Bug
> Reporter: Endre Stølsvik
> Priority: Minor
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Scenario on client:
> # Employing RedeliveryPolicy with exponential backoff (keeping maximum
> redeliveries at default 6)
> # Enabled non-blocking redelivery
> # Receiving e.g. 100 consecutive poison messages which are rolled back
> repeatedly (which eventually should DLQ after max redeliveries)
> This will result in massive redelivery delays due to a logic bug.
> The reason is that redeliveryDelay is a field variable kept on the
> ActiveMQMessageConsumer, instead of being a property on the message - or that
> the redelivery delay was calculated per message based on the redelivery count.
> When consecutive messages rollbacks multiple times, the redeliveryDelay field
> is continuously multiplied by the backoff multiplier, resulting in enormous
> delays.
> Pull Request for fix, including test:
> https://github.com/apache/activemq/pull/843
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
