I was going to say, your back or third party software. Basically, you have some processes that tried to access to the controlfiles and the controlfiles was updated at the same time, the file locked up so the DB crashed.
Printable View
I was going to say, your back or third party software. Basically, you have some processes that tried to access to the controlfiles and the controlfiles was updated at the same time, the file locked up so the DB crashed.
Are you in charge of the OS or does you shop have a sytem admin for that.Quote:
Originally Posted by Eberman
It would seem strange that something that has been running with no problems would suddenly start spitting out errors. Has anything changed on the backup for the OS?
Can you just skip the control file mount points without cancelling the entire backup?
If your controlfile is in the /u05/oracle filesystem and you're just backing up /u01/app, I doubt it. What time is the backup scheduled?
Backup Between 5 and AM. Happend this morning again. Even without OS Backup (we removed it from Scheduler)
Also once I restart Oracke everything works Ok and this control file is being updated fine. Until 5 AM?!
Eugene.
Eugene,
Its unlikely that your backup software is causing this. You don't really need to backup the Oracle control files as part of the OS backup. These are part of the DB.
This sounds like an oracle bug. What version of oracle?
I'd file a bug with Oracle.
Ken
Not really a bug.. we faced the write errors on some of the datafiles and was due to the space on that mount point being full.
Check the free space
Abhay.
Oracle version is 10.1.0.3.0
I opened SR with Oracle and they suggest to download Remote Diagnostic Agent and run and then send output report to them
E.
We are thinking along the same lines but when it happens at the same time every day it should not be free space, although it is certainly worth investigating.Quote:
Originally Posted by abhaysk
You know, it could be that they are running a batch file that is blowing up that mount point at the same time every day. Where do you have your TEMP and UNDO files located?
Cause if they swell up during a batch job, that could cause the error. Then by the time you get in they shrink down and everything looks fine.
Ahh,Quote:
Originally Posted by abhaysk
But then you'd get errors in your OS messages saying your file system was full. And you'd get a different error in your alert log. He had no OS messages.
The only space issue we found on Oracle Home drive. It was 85% full. We just cleaned up to 50%.
However control file drive has plenty of space
E.