On Mon, Dec 05, 2011 at 11:43:49 -0500, Brian Cuttler wrote:
> However I wanted to check how things worked. I'd guess that the
> data flow looks like this, with index files for the dumps.
>
> gtar -c TEE gzip (--fast) > work area > tape
> \ gzip -tf | gzip (--best) > index-file.gz
>
> Do I have that right ?
>
> Is there a switch for the compress type for the index files ?
>
On Mon, Dec 05, 2011 at 15:49:18 -0500, Brian Cuttler wrote:
> Even so, I'm sure someone will ask me again about the --best
> that shows for compression for user amanda during the nightly
> run. Was my little graphic on target ?
Amanda's internal passing-around of data is pretty crazy, but from what
I understand I believe your diagram is correct as far as how "gzip" gets
used. [I assume you meant to write "gtar -tf" at the front of the second
line.] There are definitely two branches of data being passed along, one
containing the actual tar-dump data and the other containing the index.
Of course, the gzip filter in the "dump data" branch is controlled by
the "compress" options for the dumptype of the DLE in question
(client-side v.s. server-side and "fast" v.s. "best").
The compression for the "index" branch is set up by the "dumper" program
(in the do_dump() function of dumper.c). The call there that spawns the
compression pipeline uses a hard-coded "BEST" parameter for the
compression type (so no, there is not any switch that controls that).
Nathan
----------------------------------------------------------------------------
Nathan Stratton Treadway - [email protected] - Mid-Atlantic region
Ray Ontko & Co. - Software consulting services - http://www.ontko.com/
GPG Key: http://www.ontko.com/~nathanst/gpg_key.txt ID: 1023D/ECFB6239
Key fingerprint = 6AD8 485E 20B9 5C71 231C 0C32 15F3 ADCD ECFB 6239
