-
Odd disconnection after 9.2.0.1 upgrade
Here's a strange one...
I have a database which has recently been migrated from 8.1.7 to 9.2.0.1 (part of a group of 7 databases all on Sun Solaris) which is experiencing the following problem
Code:
$ sqlplus /nolog
SQL*Plus: Release 9.2.0.1.0 - Production on Fri Mar 14 10:37:28 2003
Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.
SQL> connect / as sysdba
Connected.
SQL> SELECT
2 REQUEST_FAILURES,
3 LAST_FAILURE_SIZE,
4 MAX_FREE_SIZE,
5 FREE_SPACE,
6 REQUEST_MISSES
7 FROM
8 V$SHARED_POOL_RESERVED;
SELECT
*
ERROR at line 1:
ORA-03113: end-of-file on communication channel
In the trace file which is created the following error is given
Code:
Exception signal: 10 (SIGBUS), code: 1 (Invalid address alignment), addr: 0x40
43, PC: [0x1796a1c, kghrsc()+472]
*** 2003-03-14 10:55:13.971
ksedmp: internal or fatal error
ORA-07445: exception encountered: core dump [kghrsc()+472] [SIGBUS] [Invalid a
ddress alignment] [0x4043] [] []
At first I thought it might be a DD corruption so I have rerun catalog, catproc but still the error occurs. Of the 7 databases I migrated this is the only one experiencing this problem, I've checked Metalink and no luck, has anyone else encounter this? I really don't want to go back to 8.1.7 and I don't want to have to recreate the database again and import the data (nearly 200Gb)
Please feel free to point out my ineptitude if you spot any glaringly obvious mistakes
Cheers
Jim
Oracle Certified Professional
"Build your reputation by helping other people build theirs."
"Sarcasm may be the lowest form of wit but its still funny"
Click HERE to vist my website!
-
-
Yep, sure did
It did point to a similar problem on 8.1.7.4 but no solution was ever given.
Thanks anyway
Keep 'em coming....
Jim
Oracle Certified Professional
"Build your reputation by helping other people build theirs."
"Sarcasm may be the lowest form of wit but its still funny"
Click HERE to vist my website!
-
A bug seems to have been raised on this kghrsc function.
http://metalink.oracle.com/metalink/...G&p_id=2713652
Is it you ?
Well, the Status is "Q/A To Development "
Cheers.
P.S.: Maybe you could log an iTAR that it also occurs in 9.0.2.
-
That's not mine, but it was the one I refered to in my last post.
I've raised an iTAR to see if Oracle can shed any light it just seems a bit odd that I haven't experienced the same problem in any of the other six databases I migrated on the same box
Still Oracle can do some strange things sometimes
Regards
Jim
Oracle Certified Professional
"Build your reputation by helping other people build theirs."
"Sarcasm may be the lowest form of wit but its still funny"
Click HERE to vist my website!
-
I searched the whole world upside down (i have been searching for last 2 hrs :() but i couldn't find any thing to help you.
and do tell us what the Oracle peeps say, ( i will bet that they will ask you to upgrade to 9.2.0.2.0)
Amar
"There is a difference between knowing the path and walking the path."
-
Cheers Amar,
I don't doubt that their first response will be upgrade
After a couple of tests on other boxes I'm sometimes able to reproduce the error when I dynaically reduce the shared pool size so I think the corruption may stem from that somehow.
Still looking...
Jim
Oracle Certified Professional
"Build your reputation by helping other people build theirs."
"Sarcasm may be the lowest form of wit but its still funny"
Click HERE to vist my website!
-
Just thought I'd keep anybody who's interested up-to-date wiuth this problem.
Oracle support have suggested that cause of the problem is a heap corrution caused when the SGA was dynamically shrunk. This was backed up by the fact that later in the day the instance completely froze had had to be manually killed (kill the processes). Once the instance restarted the problem disappeared.
extract from the TAR
Explained this make sense in a way as we may have lost a memory address during the resizing and therefore the heap manager lost track of its work, thus ORA-7445 was reported.
* This could be bug 2621310, but the call stack does not match completely and I am not seeing child latches blocking each other
* Bug 2713652 looks similar, but we are not failing in strncpy, so again maybe only similar, but not the same
* I cannot say this is one or the other bug, but it is definitely some sort of heap problem
Has anyone else had these sorts of issues when changing the size of the SGA dynamincally?
Last edited by jovery; 03-17-2003 at 07:10 AM.
Jim
Oracle Certified Professional
"Build your reputation by helping other people build theirs."
"Sarcasm may be the lowest form of wit but its still funny"
Click HERE to vist my website!
-
Originally posted by jovery
Oracle support have suggested that cause of the problem is a heap corrution caused when the SGA was dynamically shrunk.
Useful piece of information. Thanks for the update.
BTW which part of SGA did you shrink, want to see if i can reproduce it at my place.
Amar
"There is a difference between knowing the path and walking the path."
-
Hi Amar,
I reduce the shared pool size by about 20%.
Jim
Oracle Certified Professional
"Build your reputation by helping other people build theirs."
"Sarcasm may be the lowest form of wit but its still funny"
Click HERE to vist my website!
Posting Permissions
- You may not post new threads
- You may not post replies
- You may not post attachments
- You may not edit your posts
-
Forum Rules
|
Click Here to Expand Forum to Full Width
|