> Eventhough we're using strict jdbc calls (i.e. non-postgreSQL
specific)
> we still get errors such as:
>
> Backend ssnen - BackendWorkerThread for backend 'pgcont1' with RAIDb
> level:1 failed (ERROR: duplicate key violates unique constraint
> "web_questions_pk")
>
> And then the sequences of the two backends become out of
> synchrionisation. We're using 2.10.8 on CentOS 4.5
Sequoia doesn't support postgres-specific features like "select
nextval('sequence')". This would be interpreted as simple select
statement (?), which will be executed on only one single backend and
therefore only on one backend the sequence value will be increased.
It is necessary to map all postgres sequence queries to the following
stored procedure syntax:
{call nextval(sequenceName)}
See Section 6.7 in the sequoia user guide.
The stored procedure call will be broadcasted to all backends, so that
the sequences on all backends will be increased.
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
