[ 
https://issues.apache.org/jira/browse/CASSANDRA-20072?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17897145#comment-17897145
 ] 

Mohammad Aburadeh commented on CASSANDRA-20072:
-----------------------------------------------

Hi [~smiklosovic] , 

1- We are still checking the network on the client side. 

     It was difficult for us to determine which client-server was causing the 
issue. I had to modify Cassandra code locally to log the client IP. Now, we are 
checking the network on that server. 

2- We open around 10-100 connections per second, each connection stays for 1-5 
seconds. 
     For the warning frequency, we get around 5 warnings every 1 minute. 

3- We have around 20 client servers. 3 of them are responsible for loading data 
into Cassandra, those servers open 10-100 connections per second. The other 
servers are for user analysis. 


Regards
Mohammad

>  Log client address when detecting unknown exception in client networking
> -------------------------------------------------------------------------
>
>                 Key: CASSANDRA-20072
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-20072
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Observability/Logging
>            Reporter: Mohammad Aburadeh
>            Assignee: Mohammad Aburadeh
>            Priority: Urgent
>             Fix For: 4.0.x, 4.1.x, 5.0.x, 5.x
>
>          Time Spent: 40m
>  Remaining Estimate: 0h
>
> We are getting hundreds of the below warnings in Cassandra logs: 
> {code:java}
> WARN  [epollEventLoopGroup-5-7] 2024-11-10 00:26:24,296 
> ExceptionHandlers.java:139 - Unknown exception in client networking
> io.netty.channel.StacklessClosedChannelException: null
>         at io.netty.channel.AbstractChannel.close(ChannelPromise)(Unknown 
> Source)
> {code}
> This means the client dropped the connection abruptly or abnormally. The 
> issue could be a network or application issue on the client side. 
> Since there could be many client servers, Cassandra must log the client IP 
> and the port so we can check what's the problem with that server. 
> I'll attached the fix for this small change. 



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to