Re: RAC clarifications

From: jason arneil <jason.arneil_at_gmail.com>
Date: Thu, 17 Dec 2009 21:19:17 +0000
Message-ID: <fda4898f0912171319l2b21a322o7d1afd6af2124fbd_at_mail.gmail.com>

2009/12/17 <debra.scarpelli_at_bell.ca>:
> Hello all, and thanks in advance for your input on these questions.
>
>
>
> For those of you who are familiar with RAC, in your experiences, how often
> are the patches (include vulnerability patches) rolling patches (i.e. no
> downtime) vs. non-rolling patches (downtime required)?

The rolling patches (including CPU's) have become much more frequent, and really the norm. However, it's not 100% and I have still encountered recently patches that are not RAC rollable. In fact i believe Oracle development now have to justify why a patch is not rollable.

Still have yet to see major upgrades be a no-downtime affair.
>
>
>
> And if anyone is using the combination of RAC with DataGuard – do you find
> this truly provides a no-downtime solution in the event of hardware failure?
> What are the drawbacks/gottchas if any?

I don't think Dataguard is sold as a no downtime solution, it can be quick to promote your standby to a primary and there are impressive claims about what is achievable with Fast-Start Failover, but I've not seen NO downtime claimed.

>
>
>
> We’re comparing a number of RDBMS solutions for a 24x365 requirement,
> including RAC. I’ve been an Oracle DBA for sometime, but not yet used RAC,
> so your experience is most appreciated!

RAC certainly does not eliminate downtime. We have application schema changes on a monthly basis on our RAC systems leading to around an hour of downtime per-month. Dataguard does not help the read-write applications here (though does enable us to keep the read-only applications going).

That's been my experience, maybe I'm doing it wrong?

jason.

--
http://jarneil.wordpress.com


>

>

>

>

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

Received on Thu Dec 17 2009 - 15:19:17 CST