Re: ORA-12571: TNS:packet writer failure

In article <t68rpllgsuqud7_at_corp.supernews.com>,   "2KFAN" <craig.wilson_at_provion.com> wrote:
> Let me add that while doing a DBShut command, I receive a ORA-3113
 error.
> I suspect this relates to the starting/stopping of the database
 yesterday
> while troubleshooting Oracle DB/Web App communication problems.
>
> I suspect I need to cycle the Oracle Process, but I am not sure how?
>
> "2KFAN" <craig.wilson_at_provion.com> wrote in message
> news:t68nu17660j339_at_corp.supernews.com...
> > Solaris 2.7, Oracle 8i
> >
> > I am having difficulty getting my Oracle Database back on-line.
> > As soon as I svrmgrl, I get the ORA-12571 error.
> > When I attempt to do anything such as connect internal or issue a
 dbshut
 or
> > dbstart command, I get a series of these errors.
> >
> > I am trying to figure out what to cycle to clear there errors.
> > Currently, cycling the entire box may be difficult since it is
 about 1000
> > miles away.
> >
> > I have root access to the box.
> >
> > NT/Netware Admin taking a heavy dose of Unix/Oracle Admin real fast.
> >
> > Any Help would be appreciated.
> >
> >
> >
>

Here's an educated guess at the problem....

Could be the base didn't die properly when you issued the dbshut and has left some IPC objects lying round. Is that base the only one on the box?
  If so, try to find the rogue semaphores / shared memory segments by issuing "ipcs -s" and "ipcs -m" respectively - if they're owned by "oracle", you can issue "ipcrm -s <sem_id>" and "ipcrm -m <shared_mem_id>" to delete them - your base should then start up.   If there's more than one base up on the box, either shut down all the others, else identify the semaphores & shared memory segments they own by doing a connect internal to each base in turn and issuing an "oradebug ipc" - the info you want will be found in trace files.

hth

Mike S

Sent via Deja.com
http://www.deja.com/ Received on Tue Jan 16 2001 - 10:33:03 CST