Re: HCC on ZFS

From: Martin Berger <martin.a.berger_at_gmail.com>
Date: Mon, 8 Aug 2016 09:55:27 +0200
Message-ID: <CALH8A93-2CdXjDOTYh0tSyWVU-U+MGh=RCa78-+GqFwv_HDwKw_at_mail.gmail.com>



Keith,

if you want to learn something about the mechanics behind, maybe this blogpost will help (for educational purpose only, of course). http://techblogsearch.com/a/hcc-on-non-exadata-how-oracle-is-detecting-storage-type.html

Martin

2016-08-03 23:43 GMT+02:00 Keith Moore <kmoore_at_zephyrus.com>:

> I was thinking of it from the other end. Hack SNMP to return the type of
> storage expected but would have to research how to do that.
>
> But like you say, not something you could do to your production (or even
> licensed) databases.
>
> Keith
>
> > Also, as I recall, someone (Randolf Geist, If I recall correctly) did a
> > blog on how to hack your DB so that the function that checks the storage
> > type always returns success, and was able to get HCC to work on any
> > arbitrary storage. Obviously, this is not something you can do to your
> > production DB, but would be a fun experiment to do on a throwaway test or
> > sandbox DB.
> >
> > -Mark
> >
> > On Wed, Aug 3, 2016 at 1:53 PM, Jonathan Lewis <
> jonathan_at_jlcomp.demon.co.uk>
> > wrote:
> >
> >>
> >> That is correct - it's purely a software choice.
> >> In fact if you have a simple Linux system you can use the
> dbms_compression
> >> package to create HCC compressed data for calculating possible
> compression
> >> factors - though you can't access the data after it's been created:
> >> http://jonathanlewis.wordpress.com/2011/10/04/hcc/
> >>
> >> Regards
> >> Jonathan Lewis
> >> http://jonathanlewis.wordpress.com
> >> _at_jloracle
> >>
> >> ________________________________________
> >> From: oracle-l-bounce_at_freelists.org [oracle-l-bounce_at_freelists.org] on
> >> behalf of Keith Moore [kmoore_at_zephyrus.com]
> >> Sent: 03 August 2016 18:37
> >> To: oracle-l_at_freelists.org
> >> Subject: HCC on ZFS
> >>
> >> We had an issue with a new Exadata/ZFS environment where a clone
> database
> >> on
> >> ZFS storage would return ORA-64307: Exadata Hybrid Columnar Compression
> is
> >> not
> >> supported for tablespaces on this storage type.
> >>
> >> That was an issue with the SNMP service on the ZFS not starting properly
> >> and
> >> has now been resolved.
> >>
> >> Based on that can someone confirm the following that seems to me to be
> >> true:
> >> 1. An Oracle database uses SNMP to verify the storage is either Exadata,
> >> ZFS
> >> or Pillar.
> >>
> >> 2. There is no technical reason HCC could not work on any storage but
> for
> >> marketing or other reasons, Oracle has decided to restrict it to only
> these
> >> platforms.
> >>
> >> Keith Moore
> >> Senior Oracle DBA - ADT Security
> >>
>

--
http://www.freelists.org/webpage/oracle-l
Received on Mon Aug 08 2016 - 09:55:27 CEST

Original text of this message