Thread: Incorrect return code for error on call of commit during 2PC

Incorrect return code for error on call of commit during 2PC

From
Ondrej Chaloupka
Date:
I'm sorry for spamming the mailing list with incomplete messages. It seems =
that my 'enter' function does not work correctly :)


it seems that PostgreSQL jdbc driver returns incorrect code during on call =
of commit during 2PC. Currently the driver returns XAER_RMERR.
I expect that it's this call
https://github.com/pgjdbc/pgjdbc/blob/master/org/postgresql/xa/PGXAConnecti=
on.java#L528

But the XA spec (http://pubs.opengroup.org/onlinepubs/009680699/toc.pdf) sa=
ys for XAER_RMERR that
"An error occurred in committing the work performed on behalf of the transa=
ction
branch and the branch=E2=80=99s work has been rolled back. Note that return=
ing this error
signals a catastrophic event to a transaction manager since other resource
managers may successfully commit their work on behalf of this branch. This =
error
should be returned only when a resource manager concludes that it can never
commit the branch and that it cannot hold the branch=E2=80=99s resources in=
 a prepared
state. Otherwise, [XA_RETRY] should be returned."

In case there is longer discussion about scenario hitting this issue and ab=
out handling this at RedHat bugzilla here:
https://bugzilla.redhat.com/show_bug.cgi?id=3D988724

From what is said here I think that PostgreSQL database should rollback the=
 transaction itself and then it could return XAException error code XAER_RM=
ERR or it should return
XAER_RMFAIL or XAER_RETRY.

Would you mind my opening a issue for PostgreSQL jdbc driver?

Thank you
Ondra Chaloupka

Re: Incorrect return code for error on call of commit during 2PC

From
Tom Lane
Date:
Ondrej Chaloupka <ochaloup@redhat.com> writes:
> it seems that PostgreSQL jdbc driver returns incorrect code during on call of commit during 2PC. Currently the driver
returnsXAER_RMERR. 

This list is mainly for reporting bugs in the core Postgres software, so
I'm not sure how closely the JDBC developers follow it.  You might have
better luck inquiring about this issue in the pgsql-jdbc list.

            regards, tom lane