Re: Backups versus snapshots
- From: "Kevin Closson" <dmarc-noreply@xxxxxxxxxxxxx> (Redacted sender "ora_kclosson@xxxxxxxxx" for DMARC)
- To: "development@xxxxxxxxxxxxxxxxx" <development@xxxxxxxxxxxxxxxxx>, "dmarc-noreply@xxxxxxxxxxxxx" <dmarc-noreply@xxxxxxxxxxxxx>, Seth Miller <sethmiller.sm@xxxxxxxxx>, "kmoore@xxxxxxxxxxxx" <kmoore@xxxxxxxxxxxx>
- Date: Thu, 18 Sep 2014 13:02:37 -0700
I'll try to be as succinct as I can. If you ever need to see your data again,
do not trade anything for backups.
________________________________
From: Martin Bach <development@xxxxxxxxxxxxxxxxx>
To: dmarc-noreply@xxxxxxxxxxxxx; Seth Miller <sethmiller.sm@xxxxxxxxx>;
"kmoore@xxxxxxxxxxxx" <kmoore@xxxxxxxxxxxx>
Cc: "andrew.kerber@xxxxxxxxx" <andrew.kerber@xxxxxxxxx>;
"Oracle-L@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx>
Sent: Thursday, September 18, 2014 12:22 PM
Subject: Re: Backups versus snapshots
Just out of interest,
What happens in that case if you have to restore the database system to what it
was last month to check if there was a data entry error? That kind of scenario,
in addition to longer term archival (think regulators) are my concerns when
thinking about snapshot only based approaches. Snapshots before software
releases, end of year processing etc sound nice to me though.
Martin
On 18 September 2014 20:46:47 CEST, Kevin Closson <dmarc-noreply@xxxxxxxxxxxxx>
wrote:
>
>array snapshots that are crash consistent don't require the database to be in
>backup mode (all EMC and most all other array snap technology I know of are
>crash consistent)
>
>________________________________
> From: Seth Miller <sethmiller.sm@xxxxxxxxx>
>To: kmoore@xxxxxxxxxxxx
>Cc: Kevin Closson <ora_kclosson@xxxxxxxxx>; "andrew.kerber@xxxxxxxxx"
><andrew.kerber@xxxxxxxxx>; "Oracle-L@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx>
>Sent: Thursday, September 18, 2014 10:06 AM
>Subject: Re: Backups versus snapshots
>
>
>
>Is it safe to assume these snapshots are being taking with the database in
>backup mode?
>
>
>What is the procedure for restoring a non-critical datafile?
>
>
>What is the procedure for repairing block corruption?
>
>
>You don't have to be an old school DBA to have concerns about this. It's not
>necessarily a bad thing to not have backups as long as everyone involved is
>aware of and accepts the consequences and loss of functionality of not having
>them.
>
>
>Seth Miller
>
>
>
>
>
>On Thu, Sep 18, 2014 at 11:46 AM, Keith Moore <kmoore@xxxxxxxxxxxx> wrote:
>
>The purpose would be to have a backup if there is a catastrophic SAN failure.
>>I know EMC and other storage vendors like to think that cannot happen but I've
>>seen it four or five times over the years, mostly due to human error.
>>
>>Your database is gone which makes the snapshot (if it still exists) invalid.
>>
>>FYI, we will be using ExtremIO and there are a lot of good things about the
>>VBlock. I'm just not comfortable with having zero backups.
>>
>>Keith
>>
>>
>>> great thread.
>>>
>>> In this specific model would the purpose of RMAN backup to make sure you're
>>> not backing up corrupt blocks?
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> Kevin Closson
>>> Chief Performance Architect - XtremIO
>>> EMC
>>>
>>>
>>> ________________________________
>>> From: Keith Moore <kmoore@xxxxxxxxxxxx>
>>> To: andrew.kerber@xxxxxxxxx
>>> Cc: "Oracle-L@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx>
>>> Sent: Thursday, September 18, 2014 9:17 AM
>>> Subject: Re: Backups versus snapshots
>>>
>>>
>>> I agree. They do not want to pay for the backup infrastructure so if that is
>>> the final decision by the architecture team, rman backups will not be an
>>> option.
>>>
>>>
>>>
>>>
>>>> Keep snapshots and rman backups. Belt and suspenders.
>>>>
>>>> Sent from my iPhone
>>>>
>>>
>>>
>>> --
>>> //www.freelists.org/webpage/oracle-l
>>
>>
>>--
>>//www.freelists.org/webpage/oracle-l
>>
>>
>>
>
>
>
--
Martin Bach
Enkitec
martincarstenbach.wordpress.com
Other related posts:
