| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> RE: very odd problem....oracle support has been no help so far.
Sheesh. This was a little over the top. I should have gone back and read
before I hit 'send'.
Oracle-l and grey goose just don't mix after a long frustrating weekend.
On a good note, I believe the situation has resulted in several reports running significantly faster this morning. I'm not sure but would imagine it's because the old db may have been missing an index.
chris
-----Original Message-----
From: Chris Stephens [mailto:ChrisStephens_at_affina.com]
Sent: Sunday, May 23, 2004 10:38 PM
To: 'oracle-l_at_freelists.org'
Subject: RE: very odd problem....oracle support has been no help so far.
Funny you should ask.
Luckily (to some extent) ..it was a reporting database that was generated from another. The database was on a machine with 3 other highly critical db's that were unaffected. I spent Friday attempting to get that db (sbcr (guess the client)) to work correctly. I've never seen anything like it. There were processes attached to the instance that root could not kill after 2 reboots and several instance restarts. When we restarted the machine I could get in but when I attempted to refresh the mv's it would hang again and eventually generate an ora-600. I worked (or didn't) with oracle support who claimed it was a known bug but the work around didn't help. I got tired of working with them (as documented by my last 42 posts) and decided that I had the weekend to recreate the database and upgrade in the process to 9204. When I attempted to deleted the db through dbassist it failed. So I went out to the mount point (we have the entire db on a single mp because I've been assured by both sa's that the san is intelligent enough to distribute the io on its own) and attempted to delete all the files. Once I issued an rm -f * the sh!t hit the fan. I couldn't ls/bdf/rm, everything was 'hanging'. At that point the sa, that vehemently denied a problem with the san from the time they had a drive failurek, all of a sudden was driving the 1hr commute back to work because 'this ain't good'. He then rebooted the server. Everything came up ok, I could 'touch' a file on the mount point and have been spending much of my weekend rebuilding the instance (and upgrading from 8172 to 9204) so that reports can run for a company that is currently on strike due to the mass movement of American labor overseas.
If much of this does not make sense, I make no apologies, I am creating my own Saturday and plan to show up tomorrow hoarse and sick pleading for comp-time on Tuesday so I can take care of the things I should have accomplished (seriously) this weekend but couldn't due to the fact that I'm the only one at the company that can do what I do after several rounds of layoffs. ...anyone work in Charlottesville, va?...i'm moving there with or without a job in September (and am headed there on wed to find a place to stay)....and am having a hard time finding employment. I am young, ignorant, energetic, and realize that I'm worth nothing more that what I can contribute to the bottom line! :) ...people at work tell me on a daily basis that I help them get their job done.
Anyways...uh...one more Saturday night!!!!
...."once in a while you get shown the light in the strangest of places if
you look at it right."
J. Garcia.........pardon the sermon! :)
-----Original Message-----
From: Sinardy Xing [mailto:SinardyXing_at_bkgcomsvc.com]
Sent: Sunday, May 23, 2004 9:49 PM
To: oracle-l_at_freelists.org
Subject: RE: very odd problem....oracle support has been no help so far.
hi Chris,
is your db ok now?
I had same (similar) problem long ago unix zombie processes got to = reboot (power off/on) Sun Solaris box, permission ok but when I tried to = "touch" a file in those mount point that contain datafile it was fail, = the reason is because my hardware (EMC) I/O Card problem. but I can see = those mount point when I do df -k, and ls -l can see the files and size.
Since you said the timestamp of those data files are freeze at certain =
point...
why don't you try to create test file in those mount points log in as db =
owner
good luck
Sinardy
-----Original Message-----
From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org]On Behalf Of Chris Stephens
Sent: 21 May 2004 23:16
To: 'oracle-l_at_freelists.org'
Subject: RE: very odd problem....oracle support has been no help so far.
Permissions are fine. Nothing has changed there.
I'm having the sysadmin run fuser now....but I believe it will only show =
the
processes that will not go away. ..i'll post the results
-----Original Message-----
From: Goulet, Dick [mailto:DGoulet_at_vicr.com]=20
Sent: Friday, May 21, 2004 9:55 AM
To: oracle-l_at_freelists.org
Subject: RE: very odd problem....oracle support has been no help so far.
Chris,
Two things come to mind,
Dick Goulet
Senior Oracle DBA
Oracle Certified 8i DBA
-----Original Message-----
From: Chris Stephens [mailto:ChrisStephens_at_affina.com]
Sent: Friday, May 21, 2004 10:40 AM
To: oracle-l_at_freelists.org
Subject: very odd problem....oracle support has been no help so far.
8.1.7.2 on HP 11
Two days ago we had a problem with our san that crashed a few of our
databases. I was told failed drives but I don't know how to validate =
=3D
that
and communication on the issue has been sketchy.
Everything seems to be fine except for one database. It is a reporting
database that contains replicated data from a remote site. =3D20
The refresh of the materialized views is scheduled for every 2 hours. =
=3D
When
the job runs we receive a 600 error with the following arguments:
ORA-00600: internal error code, arguments: [6731], [1], [0], [3], [], =
=3D
[],
[], []
ORA-00600: internal error code, arguments: [17090], [], [], [], [], [], =
=3D
[],
[]
ORA-12012: error on auto execute of job 41
ORA-12008: error in snapshot refresh path
And it gets worse. I tried shutting down the database with 'shutdown immediate'...nothing. 'Shutdown abort'..ok. Startup returns:
SQL> conn / as sysdba
Connected to an idle instance.
SQL> startup
ORACLE instance started.
Total System Global Area 87521768 bytes
Fixed Size 104936 bytes
Variable Size 38092800 bytes
Database Buffers 49152000 bytes
Redo Buffers 172032 bytes
ORA-01102: cannot mount database in EXCLUSIVE mode
SQL>
I then shutdown abort and ps -ef | grep SID. There are several =3D
processes
out there. I try to kill -9 them but they won't go away. =3D20
Sysadmins can't kill the processes either. We reboot the machine, =3D
processes
are gone. Startup the database. Run the refresh.....it 'hangs' for =3D
hours
and hours. Shutdown the database after trying to figure out what is =3D
going
on (possibly rollback from crash?...but no) and the try to restart but
receive the same unable to mount error with associated processes. The =
=3D
other
weird thing is that the majority of datafiles in the database have not =
=3D
been
touched or updated by oracle in some time. =3D20
I don't know what to do.
I opened a tar from oracle but have received no help so far. Anyone have any ideas?
Thanks
chris
-- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html ----------------------------------------------------------------- ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line. -- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html ----------------------------------------------------------------- ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line. -- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html ----------------------------------------------------------------- ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line. -- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html ----------------------------------------------------------------- ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line. -- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html ----------------------------------------------------------------- ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line. -- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------Received on Mon May 24 2004 - 08:17:17 CDT
 |
 |