Insert statements fail after 'call Nextval' in PostgreSQL cluster
-----------------------------------------------------------------
Key: SEQUOIA-975
URL: https://forge.continuent.org/jira/browse/SEQUOIA-975
Project: Sequoia
Type: Bug
Components: Core
Versions: Sequoia 2.10.9
Environment: Windows, Hibernate, PostgreSQL and Spring
Reporter: Love Singhal
I am currently facing this issue and is a blocker for me.
I have a Spring and Hibernate based Java server which is connected to
PostgreSQL database. I want to add database clustering to it using sequoia. The
server uses sequences to generate various Ids. One sequence
'hibernate_sequence' is automatically used by the Hibernate framework.
I installed a sequoia cluster of two databases with two controllers. To enable
synchronization of sequences, I changed the sql query 'select nextval' to
'{call nextval.. }' as mentioned in the documents. For hibernate, I defined the
new SQL dialect which changes the nextVal query string.
However, I am finding that the database statements that do an insert statement
along with the 'call nextval' statement in the same transaction fail
consistently. The error returned in the logs and to the server is:
> uncategorized SQLException for SQL [insert into category (name, description,
> prefix, server_path, category_id) values (?, ?, ?, ?, ?)]; SQL state [null];
> error code [0]; Batch failed for request 0: insert into category (name,
> description, prefix, server_path, category_id) values (?, ?, ?, ?, ?)
> (org.continuent.sequoia.common.exceptions.driver.DriverSQLException: Message
> of cause: write request 562949953421404 failed:
Backend ras - BackendWorkerThread for backend 'localhost2' with RAIDb level:1
failed (ERROR: current transaction is aborted, commands ignored until end of
transaction block)
The ras is the name of the virtual database.
All kinds of insert requests (not necessarily, in the category table) are
failing when they are called along with the call nextval statement. The call
nextval statement always succeeds. The error is also same for all kinds of
inserts. Everytime, backend localhost2 has some problem. I tried to locate the
cause of the problem in the logs but could not find anything that specifies the
cause (even after setting up everything to DEBUG log level). I cant do a '{call
nextval}' in a separate transaction as these calls are automatically made by
hibernate.
When I revert back the '{call nextval' statement to 'select nextval', things
work fine until some primary key gets duplicated (sequences in the two
databases become unequal). Also, the update calls along with '{call nextval'
seem to succeed.
I am now stuck with this issue and could not proceed any further with
clustering. Please help.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://forge.continuent.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
