View Revisions: Issue #564

Summary 0000564: PGPool generates error and close the conection where there is an error in the master
Revision 2019-12-09 13:21 by pengbo
Description I have pgpool-II configured only for load balancing and when there is an error in an insert in the backend0 (master) also it generates an error in the replication node (backend1) and then pgpool closes the conection. We are working with Tomcat 9, PostgreSQL 10 and in the tomcat we use the postgresql-42.2.5.jar as a driver.

THIS IS THE LOG IN THE MASTER
-----------------------------------------
< 2019-12-05 17:12:28 -05 192.168.179.7 sucursal aplivirtual > ERROR: llave duplicada viola restricción de unicidad «userrepository_pkey»
< 2019-12-05 17:12:28 -05 192.168.179.7 sucursal aplivirtual > DETALLE: Ya existe la llave (userguid, repid)=(ec002f47-989c-43bd-8c77-546a2a615b94 , 2).
< 2019-12-05 17:12:28 -05 192.168.179.7 sucursal aplivirtual > SENTENCIA: INSERT INTO gam.UserRepository(UserRepRecPwdAns, UserRepCreDate, UserRepCreUser, UserRepUpdDate, UserRepUpdUser, UserGUID, RepId, UserRepMainRoleId, UserRepSecPolId, UserRepQstUserId) VALUES($1, $2, $3, $4, $5, $6, $7, $8, $9, $10)

THIS IS THE LOG IN STANDBY
---------------------------------------------
< 2019-12-05 17:12:03 COT 192.168.179.7 sucursal aplivirtual > ERROR: no hay un savepoint con ese nombre
< 2019-12-05 17:12:03 COT 192.168.179.7 sucursal aplivirtual > SENTENCIA: ROLLBACK TO SAVEPOINT gxupdate

THIS IS PART OF CATALINA.OUT IN TOMCAT
-----------------------------
org.postgresql.util.PSQLException: FATAL: failed to read kind from backend
  Detail: kind mismatch among backends. Possible last query was: "ROLLBACK TO SAVEPOINT gxupdate" kind details are: 0[C] 1[E: no hay un savepoint con ese nombre]
  Hint: check data consistency among db nodes

We are using the PostgreSQL native replication, and the nodes are uptime correctly

Revision 2019-12-06 22:02 by eduarte
Description I have pgpool-II configured only for load balancing and when there is an error in an insert in the backend0 (master) also it generates an error in the replication node (backend1) and then pgpool closes the conection. We are working with Tomcat 9, PostgreSQL 10 and in the tomcat we use the postgresql-42.2.5.jar as a driver.

THIS IS THE LOG IN THE MASTER
-----------------------------------------
< 2019-12-05 17:12:28 -05 192.168.179.7 sucursal aplivirtual > ERROR: llave duplicada viola restricción de unicidad «userrepository_pkey»
< 2019-12-05 17:12:28 -05 192.168.179.7 sucursal aplivirtual > DETALLE: Ya existe la llave (userguid, repid)=(ec002f47-989c-43bd-8c77-546a2a615b94 , 2).
< 2019-12-05 17:12:28 -05 192.168.179.7 sucursal aplivirtual > SENTENCIA: INSERT INTO gam.UserRepository(UserRepRecPwdAns, UserRepCreDate, UserRepCreUser, UserRepUpdDate, UserRepUpdUser, UserGUID, RepId, UserRepMainRoleId, UserRepSecPolId, UserRepQstUserId) VALUES($1, $2, $3, $4, $5, $6, $7, $8, $9, $10)

THIS IS THE LOG IN STANDBY
---------------------------------------------
< 2019-12-05 17:12:03 COT 192.168.179.7 sucursal aplivirtual > ERROR: no hay un savepoint con ese nombre
< 2019-12-05 17:12:03 COT 192.168.179.7 sucursal aplivirtual > SENTENCIA: ROLLBACK TO SAVEPOINT gxupdate

THIS IS PART OF CATALINA.OUT IN TOMCAT
-----------------------------
org.postgresql.util.PSQLException: FATAL: failed to read kind from backend
  Detail: kind mismatch among backends. Possible last query was: "ROLLBACK TO SAVEPOINT gxupdate" kind details are: 0[C] 1[E: no hay un savepoint con ese nombre]
  Hint: check data consistency among db nodes

We are using the PostgreSQL native replication, and the nodes are uptime correctly