RE: Cluster Data On Tap (CDOT)

where “from” means surrounding.  

Protocols vary from site to site. To minimize and identify the archived logs required, some sites find it useful to:  

0a. alter system switch logfile

0b. record the new logfile identification

0c. back up a control file your favorite way identified as “before backup xxx”  

4a. alter system switch logfile

4b. wait for the archive log to complete

4c. back up (inclusively) the archived logs from step 0 through step 4.

4d. back up a control file your favorite way identified as “after backup xxx”  

Ls Cheng nailed it on the essentials though. Restoring from “storage” snapshots is simply physical backup recovered. And surely Ls has logfile archives and control backups down pat. Many folks rely on simply yanking the appropriate serial numbers asked for by recovery from a known location.  

Some software suites take care of all this for you. Some others effectively virtualize snapshotting to only the required database blocks.  

mwf  

From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Franky Weber Faust Sent: Tuesday, December 05, 2017 10:27 AM To: exriscer_at_gmail.com
Cc: gus.spier_at_gmail.com; oracle-l
Subject: Re: Cluster Data On Tap (CDOT)  

Just adding 0.02. :)

Remember to keep archivelogs from the begin and end backup period, you'll need them to recover when you restore your database.  

Kind regards / Cordialmente / Saludos cordiales / Sincères amitiés / Mit freundlichen Grüßen / Cordiali saluti,  

Franky Weber Faust

Oracle DBA

Skype: franky.faust

<https://docs.google.com/uc?export=download&id=1i1XxNHLsxpJw7D0yShEA0E-RlZJkyUJX&revid=0B_N3mqs0olHhTGJ1QndJOGN4S3J1OFMvakxZZTAwRFdRbGhvPQ>

<https://docs.google.com/uc?export=download&id=0B_N3mqs0olHhc0tYRkl5MlEwNlU&revid=0B_N3mqs0olHhR1JXS2tndlMwKzlTVllwQVVCTm1ud0RvWWVZPQ> <https://docs.google.com/uc?id=0B_N3mqs0olHhVlZaSVNoTC1LdnM&export=download> <https://docs.google.com/uc?export=download&id=0B_N3mqs0olHhSlZ5c0xaaXgzUUU&revid=0B_N3mqs0olHhWmFvWGJEeVJWK2k0bVVKb2MvR1JGVENjd3dJPQ> <https://docs.google.com/uc?export=download&id=0B_N3mqs0olHhcllQQTJDbmZNa0U&revid=0B_N3mqs0olHhVW9udEtVOWlnRS8rUGJFMFVkN1BKbC94aHJzPQ>
 

2017-12-05 5:12 GMT-02:00 Ls Cheng <exriscer_at_gmail.com>:

Hi

You run

1. alter database begin backup
2. storage snapshot
3. alter database end backup

so it is a consistent, recoverable backup

Thanks  

On Tue, Dec 5, 2017 at 1:44 AM, Gus Spier <gus.spier_at_gmail.com> wrote:

I am studying up on CDOT, which appears to me to be NetApp's answer to virtualization. While I work through the documentation, I am ensorcelled by the concept of database recovery by means of snapshots.  

I have never had the opportunity to try to restore/recover a database with a snapshot. To me, it sounds like the efforts of a system administrator who tried to restore and recover a production database with operating system copies of the open data files. Of course, that didn't work and much hilarity and mayhem ensured.  

I'm sure that people smarter than me have figured out how to use snapshots to recover databases, but I've never been able to reason my way through the process. Does anybody know how it is accomplished?  

Fond regards,  

Gus    

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