horizonzy opened a new issue, #3268:
URL: https://github.com/apache/bookkeeper/issues/3268

   ***Describe the bug***
   Recently, we encounter an issue. The bookie request sometimes timedout. 
   Below is the log info:
   ```
   May 11 05:11:36  pulsar[9742]: 05:11:36.586 [pulsar-io-4-8] INFO  
org.apache.bookkeeper.proto.PerChannelBookieClient - Successfully connected to 
bookie: 168.63.xx.xx:3181 [id: 0xcdfa7aed, L:/168.63.xx.xx:35012 - 
R:168.63.xx.xx/168.63.xx.xx:3181]
   May 11 05:11:36  pulsar[9742]: 05:11:36.586 [pulsar-io-4-8] INFO  
org.apache.bookkeeper.proto.PerChannelBookieClient - connection [id: 
0xcdfa7aed, L:/168.63.xx.xx:35012 - R:168.63.xx.xx/168.63.xx.xx:3181] 
authenticated as BookKeeperPrincipal{ANONYMOUS}
   May 11 06:13:43  pulsar[9742]: 06:13:43.585 
[BookKeeperClientScheduler-OrderedScheduler-0-0] INFO  
org.apache.bookkeeper.proto.PerChannelBookieClient - Timed-out 1 operations to 
channel [id: 0xcdfa7aed, L:/168.63.xx.xx:35012 - 
R:168.63.xx.xx/168.63.xx.xx:3181] for 168.xx.xx.116:3181
   May 11 06:14:18  pulsar[9742]: 06:14:18.585 
[BookKeeperClientScheduler-OrderedScheduler-0-0] INFO  
org.apache.bookkeeper.proto.PerChannelBookieClient - Timed-out 1 operations to 
channel [id: 0xcdfa7aed, L:/168.63.xx.xx:35012 - 
R:168.63.xx.xx/168.63.26.xx:3181] for 168.63.xx.xx:3181
   May 11 06:28:41  pulsar[9742]: 06:28:41.424 [pulsar-io-4-8] WARN  
org.apache.bookkeeper.proto.PerChannelBookieClient - Exception caught on:[id: 
0xcdfa7aed, L:/168.63.xx.xx:35012 - R:168.63.xx.xx/168.63.xx.xx:3181] cause: 
readAddress(..) failed: Connection timed out
   May 11 06:28:41  pulsar[9742]: 06:28:41.424 [pulsar-io-4-8] INFO  
org.apache.bookkeeper.proto.PerChannelBookieClient - Disconnected from bookie 
channel [id: 0xcdfa7aed, L:/168.63.xx.xx:35012 ! 
R:168.63.xx.xx/168.63.xx.xx:3181]
   ```
   Finally, we found that the network wall policy will drop packet after the 
connection didn't communicate with peer in a while. 
   So client packet didn't send to  server, and tcp retry reach max retry 
count, then close the connection.
   
   
   Shall we support ping-pong by interval to cover this case. 
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to