Problem when installing patchset 8.1.74 on HPUX 11 64 bit

[Henrique Silva_3]

When trying to create of replace the java system, I get this message !!!

I have increased the share_pool_size to 100 MB, and the same problem is there.

This is one node that is running master replication, and I am updating one at a time.

The reason I am installing the patch is because this error is showing up during application setup ( this is not production yet ).

Do you have any idea ? It is late and I will do some metalink's searched tomorrow, but I believe the problem might be share pool segmentation, however, there is NOTHING running on this box besides the instance :-(



SQL> create or replace java system
2 /
create or replace java system
*
ERROR at line 1:
ORA-04031: unable to allocate 4032 bytes of shared memory ("shared
pool","oracle/OracleStringSYS","joxlod: in ehe","ioc_allocate_pal")

Cheers,

Henrique Silva

[T G Manikandan]

You have mentioned that you have increased the shared pool memory to 100MB.

DId you change entries in the init.ora file.
Did you restart the instance after the change?


Else you can try doing a

sql>alter system flush shared_pool;


Revert

[Henrique Silva_3]

yes, I have changed the size on the init.ora file and restarted the instance, after verifying that there were no shared memory segments belonging to oracle via the ipcs command !!!

Henrique

[Alexander M. Ermes]

Hi there.
Have you checked your system for the latest requested patchsets on HP-UX ?
There should be a list in the release notes or in the install and upgarde guide.
What about the kernel parameters ?
They may have been enough for the old one, but are they ok after the patch ?
Rgds
Alexander M. Ermes

[T G Manikandan]

Take a look at 146599.1 doc at metalink.

Notes attached for reference.

[Andreas D. Skjervold]

Hi

Also do not forget increasing the java_pool_size to at least 50M.

Oracles java reqs:
? SHARED_POOL_SIZE >= 65 MB ? JAVA_POOL_SIZE >= 50 MB ? 50 MB free in the SYSTEM tablespace ? 250 MB of rollback segment space

Andreas

[Henrique Silva_3]

The system is got the latest UX patch sets, and java pool size was also increase, albeit to 30 MB, and my German counter part has taken over the system.

He removed replication and is running the patch install after the binaries step. He will let me know if he runs into any problems, or will pass it along to me again, once it becomes too late for him.

He is thinking about de-installing the java objects, since his app does not use them, but I told him that the Java objects update was not that time consuming. Running the cat*.sql scripts is what takes longer.

Anyways, I will keep you guys updated. I think it failed last night, for the replication was never quiesced, eventought the other instance ( master ) was down. Just a guess !!!

Thanks,

Henrique Silva

[Henrique Silva_3]

found doc 175472.1 which deals with "How to cleanup JVM in 8.1.7" and after removing and re-installing the java objects, everything seems to be fine now.

of course, had to increase large_pool_size to 100 MB and java_pool_size to 50 MB.

Thanks for the pointers.

Henrique Silva