[
https://issues.apache.org/jira/browse/TUBEMQ-99?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17105134#comment-17105134
]
Guocheng Zhang edited comment on TUBEMQ-99 at 5/12/20, 6:39 AM:
----------------------------------------------------------------
Thanks [~tiyeliyameng]
This error occurs when the thread calling getMessage () consumes data, and all
the partitions assigned to the client are in use, so there is no free partition
to pull data.
This situation generally occurs when the number of threads calling getMessage
() is greater than the number of partitions allocated by the current consumer,
such as load balancing, scaling, and the actual number of threads created is
greater than the maximum number of allocable partitions. The situation belongs
to the normal running state. It is recommended that the pull thread that calls
getMessage () sleep for 200 to 400 milliseconds when receiving such an error.
was (Author: gosonzhang):
Thanks [~tiyeliyameng]
This error indicates that when the current pull thread is consumed, all
partitions are in use, thereby telling the caller that there is no free
partition to consume.
This situation generally occurs when there are more pull threads than the
currently allocated partitions, such as load balancing, expansion and
contraction, and the actual number of threads created is greater than the
maximum number of allocable partitions. This is a normal operating state. It is
recommended to pull threads Sleep 200 ~ 400 hundred milliseconds when receiving
such error return.
> 400 - All partition in waiting, retry later!
> ---------------------------------------------
>
> Key: TUBEMQ-99
> URL: https://issues.apache.org/jira/browse/TUBEMQ-99
> Project: Apache TubeMQ
> Issue Type: Bug
> Reporter: tiyeliyameng
> Priority: Major
>
> we checked logs and found these infos，but data seemed ok
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
