Simpler Navigation coming for Servers and Operating Systems
Coming soon: a much simpler Servers and Operating Systems section of the Community. We will combine many of the older boards, and you won't have to click through so many levels to get at the information you need. If you are looking for an older board and do not find it, check the consolidated boards, as the posts are still there.
cancel
Showing results for 
Search instead for 
Did you mean: 

Oracle 8.1.6 crash

Marc Bohnert
Advisor

Oracle 8.1.6 crash

I am running HPUX 11 with Oracle 8.1.6. $GB memory with a 1GB SGA.

Oracle crashed with a segmentation violation. The Oracle error was ORA-07445. Oracle said that the online redo logs got corrupted due to a memory corruption. They do not know what triggers this. This problem is very random and Oracle suggests going to a newer release although they have not found what is causing the problem.

Has anyone experienced a similar problem.
2 REPLIES
Volker Borowski
Honored Contributor

Re: Oracle 8.1.6 crash

Hello,

several reasons for this. I would check oracle website. Even SAP-OSS has some 20 hits for various reasons for this, depending on diffrent sources.

Some are related to oracle parameters
timed_statistics
star_transformation_enabled
sql_trace

Others are OP-Sys-dependend like
PHSS_16849 core 64bit HPUX Oracle Note 73835.1 for Oracle 64-Bit on HP-UX 11.0
PHSS_16849 : s700_800 11.0 LIBCL patch should be exchanged by a superseeded one

Do not know if this helps
Volker
Jeff Schussele
Honored Contributor

Re: Oracle 8.1.6 crash

Hi Marc,

I'm always leary when vendors or developers start blaming "memory corruption" on HP-UX systems. If memory truly erred out there WILL be an entry in the syslog or the STM log. Lord knows, for the $ of HP ECC memory, it better let me know. So unless there is actually an entry, I'd turn the argument back on them, ESPECIALLY when they start the "new release" spiel. Seems awfully self-serving don't you think?
If that CAN be ruled out, AS WELL AS, faulty code - then a misconfigured kernel param would be the next likely culprit.

Rgds,
Jeff
PERSEVERANCE -- Remember, whatever does not kill you only makes you stronger!