On Fri, Dec 02, 2005 at 11:13:01AM -0700, Richard Pyne wrote:
> On 1 Dec 2005 at 8:16, Richard Pyne wrote:
> 
> > I'll give chg-zd-mtx a try.
> 
> Well, I have chg-zd-mtx working. In the process, I remembered 
> why I had chosen chg-scsi originally. The main reason is that it 
> detectes when a drive is requesting to be cleaned and will 
> initiate a clean cycle when needed rather than by access count.

I've not looked at the code, but if chg-scsi detects it via the
status request it might show up in the "mtx status" output.  If so,
then it would be 'easy' :) to add that capability to chg-zd-mtx.

> 
> It is too bad that neither ch-zd-mtx nore chg-scsi support RAIT 
> as that is the way I would like to eventially go.
> 

Amanda does support a form of it own RAIT.  But I don't think it
is well tested as few people report trying or using it.

> My desire is to set up automated backups that back up everything 
> that was changed on a daily basis and does a complete backup of 
> the production file systems at least once a week. I also want to 
> be able to do system file systems (/etc, /, /usr, etc.) on 
> demand.
> 
> Again, any configuration suggestions will be greatly 

dumpcycle 7
tapecycle "something greater than 14"
runspercycle 5 or 6 or 7

Amdump can be run anytime by hand specifying the hosts and DLEs.
You can also force, with amadmin, a level 0 of any DLEs for the
next amdump run.

A separate config could be setup sharing many of the control files
(disklist, indexes, curinfo, tapelist, changer.conf, ...) but with
record set to no and a dumpcycle of 0 (always level 0).  Then you
could run amdump for that config and leave the standard one alone.
Be careful they don't try to run at the same time.


-- 
Jon H. LaBadie                  jon AT jgcomp DOT com
 JG Computing
 4455 Province Line Road        (609) 252-0159
 Princeton, NJ  08540-4322      (609) 683-7220 (fax)