topic Re: Fatal Oracle error in Operating System - HP-UX

Fatal Oracle error

Manuel Gómez González — Mon, 13 Sep 2004 11:14:07 GMT

We have a Oracle 8.0.4 running on a HPUX 10.20 machine.

Oracle alert file noticed about a very important error reported on a trc file (attached).

Any idea about the error and how to fix it?

Re: Fatal Oracle error

Rick Garland — Mon, 13 Sep 2004 11:28:50 GMT

The dreaded ora-600 error

Have you access to metalink? Can you open a tar?

This may be the best avenue to pursue.

Re: Fatal Oracle error

Eric Antunes — Mon, 13 Sep 2004 11:28:57 GMT

Check if you are using "async" in your init.ora:

disk_async_io=?

Set it to FALSE

disk_async_io=FALSE

If you want to use "async" you'll need to create the /dev/async file...

Re: Fatal Oracle error

James A. Donovan — Mon, 13 Sep 2004 14:25:32 GMT

It appears to be a non-corruptive error. This metalink note lists possible causes.

Doc ID: Note:41472.1
Subject: ORA-600 [17034] "pin not valid"
Type: REFERENCE
Status: PUBLISHED

Content Type: TEXT/PLAIN
Creation Date: 17-NOV-1997
Last Revision Date: 06-SEP-2004

Note: For additional ORA-600 related information please read [NOTE:146580.1]

PURPOSE:

This article discusses the internal error "ORA-600 [17034]", what

it means and possible actions. The information here is only applicable

to the versions listed and is provided only for guidance.

ERROR:

ORA-600 [17034] [a] [b]

VERSIONS:

versions 7.0 to 9.2

DESCRIPTION:

When trying to unpin heaps belonging to an object we find that

the pin is not valid. This activity typically occurs when closing

cursors.

ARGUMENTS:

Arg [a] Address of the pointer to the pin

Arg [b] Pin Count

FUNCTIONALITY:

GENERIC LAYER

IMPACT:

PROCESS FAILURE

NON CORRUPTIVE - No underlying data corruption.

SUGGESTIONS:

Known Issues:

Bug# 1818676 See Note 1818676.8

Sessions exceeding PROFILE resource limits may raise OERI:17034 / OERI:17281

Fixed: 8.1.7.3, 9.0.1.2, 9.2.0.1

Bug# 1096056 See Note 1096056.8

OERI:17034 / spin after UPDATE with TRIGGER and column value too large

Fixed: 8.1.7.0

Bug# 674565 See Note 674565.8

OERI:17034 releasing a pin on a trigger in kglunp()

Fixed: 8.0.4.3, 8.0.6.2, 8.1.6.3, 8.1.7.0

Bug# 614799 See Note 614799.8

OERI:17034 possible in EXCEPTION from TRIGGER when AUDITING

Fixed: 8.0.4.2, 8.0.5.1, 8.1.5.0

Re: Fatal Oracle error

Hamdy Al-Sebaey — Mon, 13 Sep 2004 14:35:25 GMT

Hi Manuel,
ORA-00600: internal error code, arguments: [string], [string], [string], [string], [string], [string], [string], [string]

CAUSE: This is the generic internal error number for Oracle program exceptions. It indicates that a process has encountered a low-level, unexpected condition. Causes of this message include:
timeouts
file corruption
failed data checks in memory
hardware, memory, or I/O errors
incorrectly restored files
The first argument is the internal message number. Other arguments are various numbers, names, and character strings. The numbers may change meanings between different versions of Oracle.

ACTION: Report this error to Oracle Support Services after gathering the following information:

events that led up to the error
the operations that were attempted that led to the error
the conditions of the operating system and databases at the time of the error
any unusual circumstances that occurred before receiving the ORA-00600 message
contents of any trace files generated by the error
the relevant portions of the Alter files
Regards,
Hamdy

Re: Fatal Oracle error

Yogeeraj_1 — Tue, 14 Sep 2004 00:12:18 GMT

hi,

since we have little information, it can be either of these cases:
a. Related to triggers or plsql and occurring when raising an EXCEPTION from TRIGGER when AUDITING

b. Related to Triggers occurring when a pin on a trigger is being released.

c. Occurring while attempting to update a table with an invalid value (value too large for column) if the table has UPDATE TRIGGER defined. The server process may spin (appearing to hang) on disconnect.

You must contact Oracle support for more systematic help.

regards
Yogeeraj

Re: Fatal Oracle error

Manuel Gómez González — Tue, 14 Sep 2004 03:54:27 GMT

Eric,

WhatÂ´s /dev/async purpose?

Thanks everybo

Re: Fatal Oracle error

Eric Antunes — Tue, 14 Sep 2004 04:17:19 GMT

Hi Manuel,

Async IO is a most efficient form of IO available but it works better on Oracle 8i that on Oracle 8.

Try this:

- shutdown the database
- set disk_async_io=FALSE in init.ora
- startup the database

Regards,

Eric Antunes

Re: Fatal Oracle error

Eric Antunes — Tue, 14 Sep 2004 05:07:52 GMT

Manuel,

It seems they have a workaround for your issue (see Metalink Note 60692.1: http://metalink.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=NOT&p_id=60692.1):

It is possible to workaround the problem by the following means;

If the "/dev/async" device does not exists, then create it.

As the root user:

/sbin/mknod /dev/async c 101 0x0
/usr/bin/chown oracle:dba /dev/async
/usr/bin/chmod 000 /dev/async

The permissions on /dev/async of 000 ensure that it is not possible to use the async driver unintentionally.

As the oracle user, modify the "init.ora" file so that the parameters:

disk_asynch_io = false
tape_asynch_io = false

Hope this will solve your problem...

Re: Fatal Oracle error

Manuel Gómez González — Tue, 14 Sep 2004 05:19:52 GMT

Thanks Eric, weÂ´ll try and repor