Unable to start database: End-of-file error hostdef does not exit

[Former Member]

Hi,

We recently had a site failure post which we faced oracle database issues in most of our systems. There was a storage failure also. We have recovered our ECC Prd system using tape backup and are now working on dev and qas systems.

We are facing a common error in two three systems in which we are unable to even mount a database. Lisyener is running sucessfully.

When we give a startup command we get the following error :

SQL> startup

ORA-24324: service handle not initialized

ORA-01041: internal error. hostdef extension doesn't exist

SQL>

Also we get an error as below if we give a startup command :

SQL> startup mount

ORACLE instance started.

Total System Global Area 5595492352 bytes

Fixed Size                  2192488 bytes

Variable Size            1241516952 bytes

Database Buffers         4328521728 bytes

Redo Buffers               23261184 bytes

ORA-03113: end-of-file on communication channel

Process ID: 13006

Session ID: 33 Serial number: 1

SQL>

There is no update in alert log and well as user trace.

May kindly suggest some way to troubleshoot please.

[former_member182657]

HI,

ORA-01041: internal error. hostdef extension doesn't exist

For the issue you could restart the system to release allocated shared memory.

ORA-03113: end-of-file on communication channel
There is no update in alert log and well as user trace.

For this error code i would check for space on system first,there should not be stuck situations on the device.If possible share the space availability on the system first.

Second which you need to check for correct permissions & ownerships on the Database files.

Regards,

[Former Member]

Hi,

Thanks for the reply. There is adequate space on the file system and that the permissions are also correct. This issue came after a sudden hardware failure on the storage side. Please suggest some trouble shooting steps for the same.

Regards

[former_member204080]

Hi Vaibhav,

Did you try to get the file systems unmounted and mount again after hardware issue .Can you check with  unix/storage team if all LUNS are properly assigned and has the required space

Also please check if there is any trace file generated when starting oracle

Regards,

Murali

[Former Member]

Hi,

We did a couple of restarts and all the luns are available. However, the latest time stamp of trace files in the saptrace/usertrace directory are of april 10 and that of the alert.log is also of the same time. Unable to understnd the same.

Regards

[Former Member]

Sorry, forgot to add April 10 2013.