Re: GGS_DDL_SEQ on ADG env - weird (new) issue

Although, the trigger itself has this block of code which seems like it would not try to update the sequence:

IF NOT (
(dbRole = 'PRIMARY' OR dbRole = 'LOGICAL STANDBY') AND dbOpenMode = 'READ WRITE'
)
THEN
-- do not write any trace even though it should work as this is standby

                        "GGATE" .DDLReplication.setCtxInfo(-1,-1,-1,-1,-1);

RETURN; -- do not use trigger if not read/write and primary/logical_standby END IF; But then, the Oracle support says this:

*"Also, recommended to have the parameter set in a database role trigger."*

That seems to indicate I could have a database role trigger that disables system triggers on the standby. Very puzzling.

Anyone know the impact of disabling system triggers on a Physical Standby?

Chris

On Tue, Apr 30, 2019 at 10:17 AM Chris Taylor < christopherdtaylor1994_at_gmail.com> wrote:

> env: 12.1.0.2
>
> So I've got a weird issue and seems related to this doc:
>
> *ORA-00604 And ORA-16000 In ADG Due to the Internal Updates for Sequence
> GGS_DDL_SEQ (Doc ID 2503143.1) *
>
> *BACKGROUND:*
> *------------------------------------------------------------*
>
> On Friday & today our standby threw an error like this:
>
> ORA-00600: internal error code, arguments: [6817], [110546404], [], [],
> [], [], [], [], [], [], [], []
>
> 6817 related to global sequence problem.
>
> The incident file shows this:
> dbkedDefDump(): Starting incident default dumps (flags=0x2, level=3,
> mask=0x0)
> [TOC00003]
> ----- Current SQL Statement for this session (sql_id=gj5cb8uz35v3h) -----
> SELECT "GGATE" . "GGS_DDL_SEQ" .NEXTVAL FROM DUAL
>
>
> The workaround in the support document talks about disabling the trigger
> which I don't really want to do on the primary.
>
> Would it be reasonable to disable the trigger the on standby for this?
> Seems like it should be disabled on the standby anyway (or ignored) but
> doesn't seem to be happening.
>
> I'm thinking about coming up with a solution to disable the trigger on the
> standby unless that doesn't make sense?
>
> Chris
>
>
>

--
http://www.freelists.org/webpage/oracle-l