Re: Strange problem

  • From: Mladen Gogala <gogala.mladen@xxxxxxxxx>
  • To: De DBA <dedba@xxxxxxxxxx>, oracle-l <oracle-l@xxxxxxxxxxxxx>
  • Date: Mon, 1 May 2017 01:12:26 -0400

Hi Tony,

On 04/30/2017 06:49 PM, De DBA wrote:

Hi Mladen,

I've seen such errors, but on normal filesystems/ASM. They were at that time accompanied by IO errors on the underlying disk device due to failing disks.

There are no failing disks here. When the instance is restarted, no recovery is needed, everything seems to be OK.

I expect you have already ruled out hardware failure, so I would at this point investigate the ZFS volume. Has anyone changed parameters on the underlying volume, resulting in the file becoming read-only, such as setting the readonly flag on the volume?

I have no idea. I am not a Solaris expert. No readonly flag on the volume, I checked with "mount" command.

Also I suspect that merging a snapshot would change the file addressing space, and confuse the database if done with the file open. Not sure whether this is actually the case, or ZFS will handle this gracefully.



Nobody is doing snapshots on this file system.

Mladen Gogala
Oracle DBA
Tel: (347) 321-1217


Other related posts: