Re: Veritas Cluster / Oracle9i

"NetComrade" <netcomradeNSPAM_at_bookexchange.net> wrote in message news:q87762lck7uh6pi2v1ui1kbhbd03lc2uph_at_4ax.com...
> On 9 May 2006 11:19:50 -0700, "Vic" <victor.engle_at_gmail.com> wrote:
>
> >I have 2 Solaris V880s running 7 instances in 3 failover groups.
> >Usually we have 1 group on one system and the other 2 on the other. We
> >had a problem over the weekend and ended up bringing the instances up
> >outside the cluster. Also all instances and associated storage and
> >virtual IPs are currently on the same node.
> >
> >My question is this, can I simply start the cluster now on the node
> >running all the instances and just probe resources to get everything
> >back in the cluster or do I need to shutdown instances and bring them
> >up in the cluster?
>
> If you're not running RAC (know nothing about RAC or running it with
> Veritas), and every resource is really up, then just starting cluster
> (hastart) should do.. But it'd do this in off hours, something can
> always go wrong.. you might be able to start and freeze the group
> right away. Start cluster on the nodes that have the db's up first.
>
> Question.. if you have Veritas.. why did you end up bringing things up
> manually?
>
> .......
> We run Oracle 9.2.0.6 on RH4 AMD
> we are currently looking for a DBA
> remove NSPAM to email

It is often just faster to manually startup, because in my shop DBA does not have root.
Then at his liesure root can unfault the resource groups. For an HA cluster some parts of the cluster must be functional or the disks would not be mounted. Received on Fri May 12 2006 - 20:01:22 CDT