Re: RAC, ASM, and SQL-BackTrack.....

On 2006-12-06, Steve Howard <stevedhoward_at_gmail.com> wrote:
>
> JEDIDIAH wrote:
>> On 2006-12-06, Steve Howard <stevedhoward_at_gmail.com> wrote:
>> >
>> >
>> > Why do you have to archive to ASM? Can't you take the raw devices in
>> > use by ASM for archived redo logs and just create a volume group
>> > mounted as local storage?
>>
>> One big reason would be RAC. All the instances need to have
>> shared access to all of the relevant files for database recovery and
>> crash recovery.
>>
>
> This is not technically true. As long as the directory structure is
> the same on all nodes (/my_commonly_named_archivelog_destination), a
> restore processed from one node will work, restoring all threads to the

        The moment you need archive logs for threads other than the one that is doing the recovery, you need shared storage of some sort. The whole point of RAC is the whole system will tolerate the DBA going postal and taking an axe to half the servers.

> commonly named filesystem on the restoring/recovering node. I actually
> did a test case of this on RAC at one point, as we almost went in that
> direction before we decided to store them in ASM.
>
> Regards,
>
> Steve
>

-- 
NO! There are no CODICILES of Fight Club!                           |||
                                                                   / | \
That way leads to lawyers and business megacorps and credit cards!

 Posted Via Usenet.com Premium Usenet Newsgroup Services
----------------------------------------------------------
    ** SPEED ** RETENTION ** COMPLETION ** ANONYMITY **
----------------------------------------------------------        
                http://www.usenet.com