coordinator log says connection to the worker node has been closed unexpectedly

[peixubin]

When I use the COPY command to migrate large amount of data ( about 400 million rows,100gb,commit every 100000 rows ) , the coordinator and some individual woker node occasionally encounter connection exception, the worker nodes where connection exceptions occur are not fixed.

coordainator log :

2026-03-22 02:05:17.241 CST [2314407] ERROR:  connection to the remote node xxx_user@1.2.3.4:15007 failed with the following error: server closed the connection unexpectedly
                This probably means the server terminated abnormally
                before or while processing the request.
2026-03-22 02:05:17.241 CST [2314407] CONTEXT:  COPY xxxx_table, line 13381: "11096465709,2026-01-19 19:45:00,2026-01-19 19:45:57,0.0000,\N,\N,\N,\N,0.0000,220.8000,221.4000,223...."
2026-03-22 02:05:17.241 CST [2314407] STATEMENT:  COPY xxxx_table_114478 FROM STDIN DELIMITER AS ','

worker node:

2026-03-22 02:05:17.389 CST [2323216] ERROR:  canceling statement due to user request
2026-03-22 02:05:17.389 CST [2323216] CONTEXT:  COPY xxxx_table_114478, line 276
2026-03-22 02:05:17.389 CST [2323216] STATEMENT:  COPY ....... FROM STDIN WITH (format 'binary')
2026-03-22 02:05:17.389 CST [2323216] LOG:  could not send data to client: Broken pipe
2026-03-22 02:05:17.394 CST [2323216] FATAL:  connection to client lost

postgreSQL version: 18.3
citus version : 14.0