Hello,
Using
insert into schema1.x select * from x.schema2;
gives
ERROR: producer error
SQL state: XX010
the schema1.x table is replicated, has quite big XML columns. schema2.x table is distributed by hash, if it matters. Number of rows is not big, less then 100.
"select * from x.schema2" works.
Logs show repeated errors on all (four in this setup) datanodes:
/tmp/log/postgresql-2015-05-05_140709.csv:2015-05-05 15:50:57.867 CEST,"xdba","surveys",29327,"10.0.0.173:42948",5548cabf.728f,5,"BIND",2015-05-05 15:50:55 CEST,5/58,0,ERROR,XX010,"producer error",,,,,,"Remote Subplan",,"SharedQueueBind, squeue.c:565","pgxc"
/tmp/log/postgresql-2015-05-05_140709.csv:2015-05-05 15:50:57.867 CEST,"xdba","surveys",29328,"10.0.0.172:50811",5548cabf.7290,5,"BIND",2015-05-05 15:50:55 CEST,6/58,0,ERROR,XX010,"producer error",,,,,,"Remote Subplan",,"SharedQueueBind, squeue.c:565","pgxc"
/tmp/log/postgresql-2015-05-05_140709.csv:2015-05-05 15:53:39.031 CEST,"xdba","surveys",29329,"10.0.0.170:41806",5548cabf.7291,14,"BIND",2015-05-05 15:50:55 CEST,7/61,0,ERROR,XX010,"producer error",,,,,,"Remote Subplan",,"SharedQueueBind, squeue.c:565","pgxc"
/tmp/log/postgresql-2015-05-05_140709.csv:2015-05-05 15:53:39.031 CEST,"xdba","surveys",29158,"10.0.0.170:41800",5548caad.71e6,7,"REMOTE SUBPLAN",2015-05-05 15:50:37 CEST,3/301,205509,LOG,00000,"Connection error producer error",,,,,,"Remote Subplan",,"add_error_message, pgxcnode.c:2033","pgxc"
/tmp/log/postgresql-2015-05-05_140709.csv:2015-05-05 15:53:39.168 CEST,"xdba","surveys",29328,"10.0.0.172:50811",5548cabf.7290,13,"BIND",2015-05-05 15:50:55 CEST,6/61,0,ERROR,XX010,"producer error",,,,,,"Remote Subplan",,"SharedQueueBind, squeue.c:565","pgxc"
/tmp/log/postgresql-2015-05-05_140709.csv:2015-05-05 15:53:39.168 CEST,"xdba","surveys",29327,"10.0.0.173:42948",5548cabf.728f,13,"BIND",2015-05-05 15:50:55 CEST,5/61,0,ERROR,XX010,"producer error",,,,,,"Remote Subplan",,"SharedQueueBind, squeue.c:565","pgxc"
/tmp/log/postgresql-2015-05-05_140709.csv:2015-05-05 15:53:43.466 CEST,"xdba","surveys",29158,"10.0.0.170:41800",5548caad.71e6,8,"REMOTE SUBPLAN",2015-05-05 15:50:37 CEST,3/301,205509,LOG,00000,"Connection error producer error",,,,,,"Remote Subplan",,"add_error_message, pgxcnode.c:2033","pgxc"
/tmp/log/postgresql-2015-05-05_140709.csv:2015-05-05 15:53:43.467 CEST,"xdba","surveys",29158,"10.0.0.170:41800",5548caad.71e6,9,"REMOTE SUBPLAN",2015-05-05 15:50:37 CEST,3/301,205509,LOG,00000,"Connection error producer error",,,,,,"Remote Subplan",,"add_error_message, pgxcnode.c:2033","pgxc"
/tmp/log/postgresql-2015-05-05_140709.csv:2015-05-05 15:53:43.467 CEST,"xdba","surveys",29158,"10.0.0.170:41800",5548caad.71e6,10,"REMOTE SUBPLAN",2015-05-05 15:50:37 CEST,3/301,205509,ERROR,XX010,"producer error",,,,,,"Remote Subplan",,"ExecRemoteSubplan, execRemote.c:8985","pgxc"
Any ideas what could be causing this?
Master branch.
Cheers,
Krzysztof
nb. for some tables this errors appear even when the source table has no rows.
setting shared_queue_size to multiplication of the record size helped.
Still proper error should be reported and documentation improved about shared_queue_size.
We can keep this issue open until we update the documentation.