galaxysql process crashed during sysbench tests

I deployed a PolarDB-X clusters using the PXD tool，the galaxysql process crashed during sysbench tests.

/home/admin/drds-server/logs/tddl/XLog.log

2021-10-28 17:14:00.403 ERROR ERR-CODE: [PXC-10004][ERR_X_PROTOCOL_RESULT] Fatal error when fetch data: Expectation failed: no_error HY000
com.alibaba.polardbx.common.exception.TddlRuntimeException: ERR-CODE: [PXC-10004][ERR_X_PROTOCOL_RESULT] Fatal error when fetch data: Expectation failed: no_error HY000
at com.alibaba.polardbx.rpc.result.XResult.internalFetchOneObject(XResult.java:921)
at com.alibaba.polardbx.rpc.result.XResult.waitFinish(XResult.java:508)
at com.alibaba.polardbx.rpc.client.XSession.flushIgnorable(XSession.java:889)
at com.alibaba.polardbx.rpc.pool.XConnection.close(XConnection.java:112)
at com.alibaba.polardbx.gms.metadb.MetaDbDataSource.initTsoServicesX(MetaDbDataSource.java:132)
at com.alibaba.polardbx.gms.metadb.MetaDbDataSource.initXDataSourceByJdbcProps(MetaDbDataSource.java:170)
at com.alibaba.polardbx.gms.metadb.MetaDbDataSource.initMetaDbDataSource(MetaDbDataSource.java:181)
at com.alibaba.polardbx.gms.metadb.MetaDbDataSource.doInit(MetaDbDataSource.java:121)
at com.alibaba.polardbx.common.model.lifecycle.AbstractLifecycle.init(AbstractLifecycle.java:43)
at com.alibaba.polardbx.gms.metadb.MetaDbDataSource.getInstance(MetaDbDataSource.java:462)
at com.alibaba.polardbx.gms.metadb.MetaDbDataSource.initMetaDbDataSource(MetaDbDataSource.java:488)
at com.alibaba.polardbx.config.loader.ServerLoader.initPolarDbXComponents(ServerLoader.java:207)
at com.alibaba.polardbx.config.loader.ServerLoader.load(ServerLoader.java:181)
at com.alibaba.polardbx.config.loader.ServerLoader.doInit(ServerLoader.java:91)
at com.alibaba.polardbx.common.model.lifecycle.AbstractLifecycle.init(AbstractLifecycle.java:43)
at com.alibaba.polardbx.CobarConfig.initCobarConfig(CobarConfig.java:68)
at com.alibaba.polardbx.CobarConfig.(CobarConfig.java:61)
at com.alibaba.polardbx.CobarServer.(CobarServer.java:141)
at com.alibaba.polardbx.CobarServer.(CobarServer.java:95)
at com.alibaba.polardbx.server.TddlLauncher.main(TddlLauncher.java:82)

Start with /home/admin/drds-server/logs/tddl/XLog.log and trace the failure through XResult.internalFetchOneObject, XSession.flushIgnorable, and MetaDbDataSource.initTsoServicesX. Reproduce the crash during sysbench tests with the PXD-deployed cluster; done means the galaxysql process no longer crashes under that workload.