[
https://issues.apache.org/jira/browse/KAFKA-7765?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
huxihx updated KAFKA-7765:
--------------------------
Description:
Currently, controller creates sockets for every living brokers without idle
timeout. However, processor threads from other brokers still could close these
sockets if no requests flow through them within `connections.max.idle.ms`.
Lots of CLOSE_WAITs were left when those sockets were closed by remote peer
since controller's RequestSendThread will not check if they are closed by peer.
I think we need to figure out a way to record which channels should be
maintained and have them excluded by IdleExpiryManager. A naive method is to
augment KafkaChannel, making it have a field indicating whether this channel
should be kept alive.
Does it make any sense?
was:
Currently, controller creates sockets for every living brokers without idle
timeout. However, other brokers' processor threads still could close these
sockets if no requests flow through them within `connections.max.idle.ms`.
Lots of CLOSE_WAITs were left when those sockets were closed by remote peer
since controller's RequestSendThread will not check if they are closed by peer.
I think we need to figure out a way to record which channels should be
maintained and have them excluded by IdleExpiryManager. A naive method is to
augment KafkaChannel, making it have a field indicating whether this channel
should be kept alive.
Does it make any sense?
> IdleExpiryManager should not passively close socket used by controller
> ----------------------------------------------------------------------
>
> Key: KAFKA-7765
> URL: https://issues.apache.org/jira/browse/KAFKA-7765
> Project: Kafka
> Issue Type: Bug
> Components: network
> Affects Versions: 2.1.0
> Reporter: huxihx
> Priority: Major
>
> Currently, controller creates sockets for every living brokers without idle
> timeout. However, processor threads from other brokers still could close
> these sockets if no requests flow through them within
> `connections.max.idle.ms`.
> Lots of CLOSE_WAITs were left when those sockets were closed by remote peer
> since controller's RequestSendThread will not check if they are closed by
> peer.
> I think we need to figure out a way to record which channels should be
> maintained and have them excluded by IdleExpiryManager. A naive method is to
> augment KafkaChannel, making it have a field indicating whether this channel
> should be kept alive.
> Does it make any sense?
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
