[
https://issues.apache.org/jira/browse/IGNITE-22784?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Roman Puchkovskiy updated IGNITE-22784:
---------------------------------------
Description:
Logs like the following
[2024-07-19T18:36:19,849][WARN
][%ipact_n_1%MessagingService-inbound-0-0][TrackableNetworkMessageHandler]
Message handling has been too long [duration=10ms,
message=[AppendEntriesRequestImpl [committedIndex=1,
data=java.nio.HeapByteBuffer[pos=0 lim=0 cap=0], entriesList=null,
groupId=metastorage_group, peerId=ipact_n_1, prevLogIndex=1, prevLogTerm=1,
serverId=ipact_n_0, term=1, timestamp=HybridTimestamp [physical=2024-07-19
18:36:19:825 +0400, logical=0, composite=112813655970611200]]]]
and
[2024-07-19T18:36:19,861][WARN
][%ipact_n_0%MessagingService-inbound-0-0][TrackableNetworkMessageHandler]
Message handling has been too long [duration=7ms,
message=[ReadActionRequestImpl [command=GetCurrentRevisionCommandImpl [],
groupId=metastorage_group, readOnlySafe=true]]]
can be seen in logs (in this case the test is ItPublicApiColocationTest).
It can be seen that handling these messages in an inbound thread takes more
than 5ms, but such handling should only dispatch a request to the corresponding
executor, it should not take so much time.
> Speed up AppendEntriesRequest and ReadActionRequest processing on inbound
> thread
> --------------------------------------------------------------------------------
>
> Key: IGNITE-22784
> URL: https://issues.apache.org/jira/browse/IGNITE-22784
> Project: Ignite
> Issue Type: Improvement
> Reporter: Roman Puchkovskiy
> Priority: Major
> Labels: ignite-3, long-message-handling, threading
>
> Logs like the following
> [2024-07-19T18:36:19,849][WARN
> ][%ipact_n_1%MessagingService-inbound-0-0][TrackableNetworkMessageHandler]
> Message handling has been too long [duration=10ms,
> message=[AppendEntriesRequestImpl [committedIndex=1,
> data=java.nio.HeapByteBuffer[pos=0 lim=0 cap=0], entriesList=null,
> groupId=metastorage_group, peerId=ipact_n_1, prevLogIndex=1, prevLogTerm=1,
> serverId=ipact_n_0, term=1, timestamp=HybridTimestamp [physical=2024-07-19
> 18:36:19:825 +0400, logical=0, composite=112813655970611200]]]]
> and
> [2024-07-19T18:36:19,861][WARN
> ][%ipact_n_0%MessagingService-inbound-0-0][TrackableNetworkMessageHandler]
> Message handling has been too long [duration=7ms,
> message=[ReadActionRequestImpl [command=GetCurrentRevisionCommandImpl [],
> groupId=metastorage_group, readOnlySafe=true]]]
> can be seen in logs (in this case the test is ItPublicApiColocationTest).
> It can be seen that handling these messages in an inbound thread takes more
> than 5ms, but such handling should only dispatch a request to the
> corresponding executor, it should not take so much time.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
