Re: 11gR2 rman compression algorithms
- From: Seth Miller <sethmiller.sm@xxxxxxxxx>
- To: Jeremy Schneider <jeremy.schneider@xxxxxxxxxxxxxx>
- Date: Thu, 4 Sep 2014 12:22:58 -0500
This is from the Oracle University 11g Admin course.
*The following level or compression ratios are available:• LOW: This level
is the fastest. It provides less compression than MEDIUM, but uses the
least CPU. (It corresponds to the LZO compression.)• MEDIUM: This level
provides a good balance of CPU usage and compression ratio. (It corresponds
to the ZLIB compression.)• HIGH: This level provides the best compression
ratio, but consumes the most CPU. (It corresponds to the GZIP
compression.)• BASIC: This corresponds to BZIP2 (10g style compression).*
Seth Miller
On Thu, Sep 4, 2014 at 11:45 AM, Jeremy Schneider <
jeremy.schneider@xxxxxxxxxxxxxx> wrote:
> In 11gR2, Oracle added some new rman compression algorithms (extra
> licensing required). They also changed the semantics - calling them low,
> medium and high.
>
> I'm having a bad search day... I've been searching the docs and the MOS
> knowledge-base but I haven't found any hints whatsoever about the
> underlying algorithms. The internet seems full of blog posts claiming that
> low is LZO, medium is the old ZLIB, and high is a modified BZ2. But I
> haven't yet found a blog post that has any kind of real citation or even a
> test case to confirm this. (How hard would it be to pstack an in-progress
> backup or recovery and check function names?)
>
> I really hate it when the internet is full of stupid blog posts that make
> claims but don't substantiate them. Especially if the proof
> isn't anywhere in public.
>
> Does anyone know of an actual doc/mos reference or public test case that
> substantiates the underlying algorithms?
>
> Probably I'm just missing some obvious link.
>
> -Jeremy
>
> --
> http://about.me/jeremy_schneider
>
Other related posts:
