Re: Issues with server and need to restore controlfiles and spfiles

  • From: Tim Gorman <tim.evdbt@xxxxxxxxx>
  • To: oracle-l@xxxxxxxxxxxxx
  • Date: Mon, 10 Jan 2022 12:34:26 -0800

Sandy,

Just curious:  which mechanism are you using to persist the addresses of the SCSI devices across reboots?  ASMLIB?  Linux UDEV?

If the pathname of the ASM disk are something like "/dev/sdc1" or "/dev/dm-3", then you're using device naming which changes on reboot, and probably should reviewOracle Support Document 1365511.1 (How to Configure LUNs for ASM Disks using WWID, DM-Multipathing, and ASMLIB on RHEL 5/OL 5 and  RHEL 6/OL 6  and RHEL7) which can be found at: https://support.oracle.com/epmos/faces/DocumentDisplay?id=1365511.1.

And, as a side note, please press them to upgrade to at least to 11.2.0.4 on RHEL7?  Friends don't let friends run 11.2.0.1, as 11.2.0.1 is an example of where the maxim "/if it ain't broke don't fix it/" never even applied;  it was broke from the start.  The only good 11g database is an 11.2.0.4 database;  every other release in 11.1 and 11.2 is a crisis waiting to happen.

Hope this helps...

-Tim


On 1/10/2022 11:36 AM, Sandra Becker wrote:

Oracle Database & Grid:  11.2.0.1
OS:  RHEL5.5

We had some issues with a database server and the 9 databases stopped working.  The sys admins suggested a reboot, which we did.  GRID/ASM would not come back online after the reboot.  It says it can't find the ASM disks, but everything the sys admins look at say they are there and online.

We tried restoring the  $GRID_HOME, but that didn't change anything.  Our Rimini support said a lot of libraries were missing from the $GRID_HOME, so another DBA tried to getting the missing files.  Still no luck.  It's a two-node RAC, but only one of the nodes was functional at the time we rebooted the first node.  The second has the same issues as the first, so we can't bring anything online there either.

We tried setting up on another comparable cluster, but due to the way the database configurations were done in CommVault for the "bad" server, CommVault thinks they are single instance and won't let us restore to a RAC.

Ultimately, we want to move these databases to a non-ASM server in AWS.  Any suggestions as to what we can do?  I was wondering if we could reinstall GRID/ASM on the "bad" server and then use Commvault to do our restores.  Is that totally off base or a viable option?

--
Sandy B.

Other related posts: