[jira] [Comment Edited] (FLINK-35151) Flink mysql cdc will stuck when suspend binlog split and ChangeEventQueue is full
[ https://issues.apache.org/jira/browse/FLINK-35151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17838452#comment-17838452 ] Xin Gong edited comment on FLINK-35151 at 4/18/24 4:46 AM: --- I get an idea to address this issue by setting "currentTaskRunning || queue.remainingCapacity() == 0" for BinlogSplitReader#pollSplitRecords. [~Leonard] PTAL was (Author: JIRAUSER292212): I get an idea to address this issue by set "currentTaskRunning || queue.remainingCapacity() == 0" for BinlogSplitReader#pollSplitRecords. [~Leonard] PTAL > Flink mysql cdc will stuck when suspend binlog split and ChangeEventQueue is > full > -- > > Key: FLINK-35151 > URL: https://issues.apache.org/jira/browse/FLINK-35151 > Project: Flink > Issue Type: Bug > Components: Flink CDC > Environment: I use master branch reproduce it. >Reporter: Xin Gong >Priority: Major > Attachments: dumpstack.txt > > > Flink mysql cdc will stuck when suspend binlog split and ChangeEventQueue is > full. > Reason is that producing binlog is too fast. > MySqlSplitReader#suspendBinlogReaderIfNeed will execute > BinlogSplitReader#stopBinlogReadTask to set > currentTaskRunning to be false after MysqSourceReader receives binlog split > update event. > MySqlSplitReader#pollSplitRecords is executed and > dataIt is null to execute closeBinlogReader when currentReader is > BinlogSplitReader. closeBinlogReader will execute > statefulTaskContext.getBinaryLogClient().disconnect(), it could dead lock. > Because BinaryLogClient#connectLock is not release when > MySqlStreamingChangeEventSource add element to full queue. > > You can set StatefulTaskContext#queue to be 1 and run UT > NewlyAddedTableITCase#testRemoveAndAddNewTable. > -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Comment Edited] (FLINK-35151) Flink mysql cdc will stuck when suspend binlog split and ChangeEventQueue is full
[ https://issues.apache.org/jira/browse/FLINK-35151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17838452#comment-17838452 ] Xin Gong edited comment on FLINK-35151 at 4/18/24 4:44 AM: --- I get an idea to address this issue by set "currentTaskRunning || queue.remainingCapacity() == 0" for BinlogSplitReader#pollSplitRecords. [~Leonard] PTAL was (Author: JIRAUSER292212): I get an idea to address this issue by set currentTaskRunning || queue.remainingCapacity() == 0 for BinlogSplitReader#pollSplitRecords. [~Leonard] PTAL > Flink mysql cdc will stuck when suspend binlog split and ChangeEventQueue is > full > -- > > Key: FLINK-35151 > URL: https://issues.apache.org/jira/browse/FLINK-35151 > Project: Flink > Issue Type: Bug > Components: Flink CDC > Environment: I use master branch reproduce it. >Reporter: Xin Gong >Priority: Major > Attachments: dumpstack.txt > > > Flink mysql cdc will stuck when suspend binlog split and ChangeEventQueue is > full. > Reason is that producing binlog is too fast. > MySqlSplitReader#suspendBinlogReaderIfNeed will execute > BinlogSplitReader#stopBinlogReadTask to set > currentTaskRunning to be false after MysqSourceReader receives binlog split > update event. > MySqlSplitReader#pollSplitRecords is executed and > dataIt is null to execute closeBinlogReader when currentReader is > BinlogSplitReader. closeBinlogReader will execute > statefulTaskContext.getBinaryLogClient().disconnect(), it could dead lock. > Because BinaryLogClient#connectLock is not release when > MySqlStreamingChangeEventSource add element to full queue. > > You can set StatefulTaskContext#queue to be 1 and run UT > NewlyAddedTableITCase#testRemoveAndAddNewTable. > -- This message was sent by Atlassian Jira (v8.20.10#820010)
[jira] [Comment Edited] (FLINK-35151) Flink mysql cdc will stuck when suspend binlog split and ChangeEventQueue is full
[ https://issues.apache.org/jira/browse/FLINK-35151?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17838452#comment-17838452 ] Xin Gong edited comment on FLINK-35151 at 4/18/24 4:43 AM: --- I get an idea to address this issue by set currentTaskRunning || queue.remainingCapacity() == 0 for BinlogSplitReader#pollSplitRecords. [~Leonard] PTAL was (Author: JIRAUSER292212): I get an idea to address this issue by set currentTaskRunning || queue.remainingCapacity() == 0 for BinlogSplitReader#pollSplitRecords. > Flink mysql cdc will stuck when suspend binlog split and ChangeEventQueue is > full > -- > > Key: FLINK-35151 > URL: https://issues.apache.org/jira/browse/FLINK-35151 > Project: Flink > Issue Type: Bug > Components: Flink CDC > Environment: I use master branch reproduce it. >Reporter: Xin Gong >Priority: Major > Attachments: dumpstack.txt > > > Flink mysql cdc will stuck when suspend binlog split and ChangeEventQueue is > full. > Reason is that producing binlog is too fast. > MySqlSplitReader#suspendBinlogReaderIfNeed will execute > BinlogSplitReader#stopBinlogReadTask to set > currentTaskRunning to be false after MysqSourceReader receives binlog split > update event. > MySqlSplitReader#pollSplitRecords is executed and > dataIt is null to execute closeBinlogReader when currentReader is > BinlogSplitReader. closeBinlogReader will execute > statefulTaskContext.getBinaryLogClient().disconnect(), it could dead lock. > Because BinaryLogClient#connectLock is not release when > MySqlStreamingChangeEventSource add element to full queue. > > You can set StatefulTaskContext#queue to be 1 and run UT > NewlyAddedTableITCase#testRemoveAndAddNewTable. > -- This message was sent by Atlassian Jira (v8.20.10#820010)
