Re: Data Guard: ORA-16534: no more requests accepted

Actually, the standby server is monitoring on 1522 and the primary on 1526 (just to keep us on our toes). What I figured out was that I had to bounce the new primary instance and that cleared the queue. The process we had to go through was half automatic (switchover to <physical standby name>) and half manual. But we were able to move eight 10g databases from the RAC to the standby. If anyone ever finds themselves in that situation, I wrote up the steps we followed. They might not be the best or most efficient but they worked. Thank you for responding Rajeev! Unfortunately, these things always seem to happen on the weekend when happy DBAs are sleeping.

Kevin Lidh

On Sun, 2007-09-16 at 17:40 -0400, Rajeev Prabhakar wrote:
> Kevin
>
> Have you already tried the non-default listener(say - on port 1522 )
> option so
> as to NOT use automatic service registration ?
>
> Rajeev
>
> On 9/16/07, Kevin Lidh <kevin.lidh_at_gmail.com> wrote:
> We have an environment Solaris (SPARC) 9 with Oracle
> 10.2.0.1.0 with a
> two-node RAC database. We used Grid Control to create a
> standby on
> another node and it was successful. We then tried to do a
> switchover
> using DGMGRL. The new primary was up and successful but the
> standby
> opened on node 2 and logs weren't being shipped. I shutdown
> the standby
> on node 2 and started the database manually on node 1 with the
>
> 'alter database recover managed standby database through all
> switchover
> disconnect using current logfile'
>
> All of a sudden logs were being shipped and applied. Then CRS
> started
> the instance on node 2 and all was good.
>
> I then angered the DG gods by attempting to switch back from
> the 1-node
> primary back to the 2-node standby. If I do a "verify
> configuration"
> from Grid Control, it gives me an ORA-16534 saying it won't
> accept
> anymore requests. The "action" specified is to wait but
> nothing is
> happening. Metalink has one example of this error but it was
> when a
> listener wasn't defined correctly and the switchover failed.
>
> What I'm wondering is, is there any way to clear the DG queue
> and
> attempt the switchover again?
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

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