Umesh Kumar Kumawat created HBASE-29714:
-------------------------------------------
Summary: DEFAULT_RS_REMOTE_PROC_RETRY_LIMIT is too low causing RS
abortion in just 3 seconds
Key: HBASE-29714
URL: https://issues.apache.org/jira/browse/HBASE-29714
Project: HBase
Issue Type: Improvement
Components: proc-v2
Reporter: Umesh Kumar Kumawat
Fix For: 3.0.0, 2.6.5
DEFAULT_RS_RPC_RETRY_INTERVAL is 100 second and
DEFAULT_RS_REMOTE_PROC_RETRY_LIMIT is 5 so we are aborting the RS in just 3
seconds
2025-11-07 11:48:36,298 WARN [RSProcedureDispatcher-pool-30146]
procedure.RSProcedureDispatcher - request to rs1,xxx,1762426870165 failed due
to org.apache.hadoop.hbase.exceptions.ConnectionClosedException: Call to
address=rs1:xxx failed on local exception:
org.apache.hadoop.hbase.exceptions.ConnectionClosedException: Connection
closed, try=0, retrying...
2025-11-07 11:48:36,399 WARN [RSProcedureDispatcher-pool-30136]
procedure.RSProcedureDispatcher - request to rs1,xxx,1762426870165 failed due
to java.io.IOException: Call to address=rs1:xxx failed on local exception:
java.io.IOException: Can not send request because relogin is in progress.,
try=1, retrying... ,
2025-11-07 11:48:36,799 WARN [RSProcedureDispatcher-pool-30116]
procedure.RSProcedureDispatcher - request to rs1,xxx,1762426870165 failed due
to java.io.IOException: Call to address=rs1:xxx failed on local exception:
java.io.IOException: Can not send request because relogin is in progress.,
try=2, retrying...
2025-11-07 11:48:37,700 WARN [RSProcedureDispatcher-pool-30088]
procedure.RSProcedureDispatcher - request to rs1,xxx,1762426870165 failed due
to java.io.IOException: Call to address=rs1:xxx failed on local exception:
java.io.IOException: Can not send request because relogin is in progress.,
try=3, retrying...
2025-11-07 11:48:39,301 WARN [RSProcedureDispatcher-pool-30124]
procedure.RSProcedureDispatcher - Number of retries 5 exceeded limit 5 for the
given error type. Scheduling server crash for rs1,xxx,1762426870165
as per the logs in just 100(1^2+2^2+3^2+4^2) seconds RS got aborted. Taking hte
example of RITThreshold as 60 seconds, we should at least wait for 30 seconds.
that will happen if we wait keep the DEFAULT_RS_REMOTE_PROC_RETRY_LIMIT 10
(28.5 seconds).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
