Re: Urgent!!

  • From: Kjetil Strønen <kjetil@xxxxxxxxxx>
  • To: "Oracle-L@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx>
  • Date: Tue, 27 Apr 2010 12:42:05 +0200

On Tue, 2010-04-27 at 20:13 +1000, Nuno Souto wrote:
> Ozgur Ozdemircili wrote,on my timestamp of 27/04/2010 2:58 AM:
> >  Things do not get lost without any trace. What is very suspiciuos here 
> > is whole partition getting lost. If it was the storm causing it you`d 
> > probably end up with a fried hdd/ power supply.
> > 
> > Just to be on the safe side, get a second opinion.Check /etc/fstab to 
> > make sure everything is mounted. Check history and last commands.
> > 
> 
> Yup, that's be my approach as well. Highly suss
> how it all vanished.  "Mistery" things like that
> just don't happen, they are urban myth.
> I'd look at it from the angle of a hack.

The one time I've witnessed a similar incident, it turned out to be the
result of multiple "someone"'s unintended blunders. A RAC installation
had been moved to new servers, but the SAN admins hadn't revoked access
to the disks from the old HBA's. When the old servers where put in use
again, for other purposes, the sysadmin had merrily formatted an LVM
volume on all accessible disks. ASM doesn't play well when it's disks
suddenly get a header saying they're fresh LVM volumes created by
"localhost.localdomain"...

("dd if=<device> bs=1M count=10|hexdump -C|less" gave it away)

Spent 24 hours with support, trying to recover the ASM disks, but they
were too mangled, so we had to recreate the ASM diskgroups, and restore.

--Kjetil

--
//www.freelists.org/webpage/oracle-l


Other related posts: