[jira] [Commented] (KAFKA-5033) Reconsider default retries for idempotent producer
[ https://issues.apache.org/jira/browse/KAFKA-5033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16016381#comment-16016381 ] ASF GitHub Bot commented on KAFKA-5033: --- Github user asfgit closed the pull request at: https://github.com/apache/kafka/pull/3091 > Reconsider default retries for idempotent producer > -- > > Key: KAFKA-5033 > URL: https://issues.apache.org/jira/browse/KAFKA-5033 > Project: Kafka > Issue Type: Sub-task > Components: clients, core, producer >Reporter: Ismael Juma >Assignee: Apurva Mehta >Priority: Blocker > Labels: exactly-once > Fix For: 0.11.0.0 > > > We currently set the default to 3 if idempotence is enabled. There was a > brief discussion in the PR, but it would be good to explain this choice in > more detail. Namely, what cases are we trying to cover with this number and > have we considered the fact that with the default retry backoff of 100ms, 3 > retries are used pretty quickly. > Is there a downside in using a larger retry count? > cc [~apurva] -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (KAFKA-5033) Reconsider default retries for idempotent producer
[ https://issues.apache.org/jira/browse/KAFKA-5033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16016240#comment-16016240 ] ASF GitHub Bot commented on KAFKA-5033: --- GitHub user apurvam opened a pull request: https://github.com/apache/kafka/pull/3091 KAFKA-5033: Set default retries for the idempotent producer to be infinite. You can merge this pull request into a Git repository by running: $ git pull https://github.com/apurvam/kafka KAFKA-5033-bump-retries-for-idempotent-producer Alternatively you can review and apply these changes as the patch at: https://github.com/apache/kafka/pull/3091.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #3091 > Reconsider default retries for idempotent producer > -- > > Key: KAFKA-5033 > URL: https://issues.apache.org/jira/browse/KAFKA-5033 > Project: Kafka > Issue Type: Sub-task > Components: clients, core, producer >Reporter: Ismael Juma >Assignee: Apurva Mehta >Priority: Blocker > Labels: exactly-once > Fix For: 0.11.0.0 > > > We currently set the default to 3 if idempotence is enabled. There was a > brief discussion in the PR, but it would be good to explain this choice in > more detail. Namely, what cases are we trying to cover with this number and > have we considered the fact that with the default retry backoff of 100ms, 3 > retries are used pretty quickly. > Is there a downside in using a larger retry count? > cc [~apurva] -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (KAFKA-5033) Reconsider default retries for idempotent producer
[ https://issues.apache.org/jira/browse/KAFKA-5033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16016152#comment-16016152 ] Apurva Mehta commented on KAFKA-5033: - A default of `MAX_INT` seems reasonable here. > Reconsider default retries for idempotent producer > -- > > Key: KAFKA-5033 > URL: https://issues.apache.org/jira/browse/KAFKA-5033 > Project: Kafka > Issue Type: Sub-task > Components: clients, core, producer >Reporter: Ismael Juma >Assignee: Apurva Mehta > Labels: exactly-once > Fix For: 0.11.0.0 > > > We currently set the default to 3 if idempotence is enabled. There was a > brief discussion in the PR, but it would be good to explain this choice in > more detail. Namely, what cases are we trying to cover with this number and > have we considered the fact that with the default retry backoff of 100ms, 3 > retries are used pretty quickly. > Is there a downside in using a larger retry count? > cc [~apurva] -- This message was sent by Atlassian JIRA (v6.3.15#6346)