Eugene,
what is your hardware configuration on the disk where the control file is stored?
What is you fs type and mounting options?
Printable View
Eugene,
what is your hardware configuration on the disk where the control file is stored?
What is you fs type and mounting options?
Well any n/w outages at 5am daily?.. and that ur control file is remote to the server?
Hardware config:
aix- jsf2
all oracle filessystems are mirroed
/dev/u03lv -- /u03 jfs2 125829120 rbr,rbw,dio,rw no no
Also no network issues.
Would UNIX Server reboot solve this issue? We had not reboot this box for long time
E.
A reboot probably won't help, but it wouldn't hurt either. Plus if there are hardware issues the startup diagnostics should help pick these up. The startup FSCK should also pick up any file system issues. Its worth a try.
What is the storage device? How is it attached? IS it an EMC or some other SAN?
Can you relocate control03 to say an internal disk (temporarily)?
Just a thought. I wonder if you're getting intermittent write errors on the mirrored device? hardware mirroring or software?
All I know its an ibm fast-t900
We will reboot Unix Server at night time and I will move control files just like suggested yesterday. See if it helps us.
Eugene.
Eugene..
I looked. The hardware supports internal mirroring. It is a FCAL (fibre chanel arbitrated loop) device. There are basically two connections to each physical drive on seperate loops.
I'd get you sys admin too look and see if there are any hardware errors and run diagnostics on this as well during your reboot/downtime.
Is this located on site? I once had a strange problem with db crashing. The cleaning personel were unplugging the disk subsystem to plug in their vacum.
But, why the crash is happeneing only at 5AM..? If there is any general H/w issue, the DB can crash at any point not restricted to 5AM?
Last nigth we stop Unix server OS backup and stop RMAN backup as well. No downtime!!
We will bring back RMAN only to see if this Unix Server OS backup casing trouble.
Eugene
There You go... Its one of those then. Process of elimination from here out.
Let us know which one caused the problem.
Ken
first mount the db and issue an recover automatic database may solve the problem if not recreate the controlfile ..I think you know the rest.