On Wed, Mar 15, 2006 at 04:08:05PM +0100, Thomas Widhalm wrote:
> On Wed, 2006-03-15 at 09:45 -0500, Jon LaBadie wrote:
> > On Wed, Mar 15, 2006 at 01:03:51PM +0100, Thomas Widhalm wrote:
> > > 
> > > Here are the 2 parts of the config:
> > > 
> > 
> > I took the liberty of deleting the comments and unneeded
> > tapetypes/dumptypes.  800 lines dropped to about 50.
> > Hope I did not drop anything significant to the discussion.
> 
> So sorry. We had a lot of troubles migrating our other servers from
> legato to data protector, so I missed to clean the files before posting.
> I'm really ver sorry.

The amanda.conf file, as distributed, is heavily overloaded
with commentary and unused definitions.  It seldom is cleaned up.
I was not admonishing your posting, just explaining what I did
in case there were errors in my editing.

> > Amazing how "simple" amanda.conf can become.
> > 
> > I put two questions inline.
> > 
> > || 
> > || Here are the 2 parts of the config:
> > || 
> > || first file:
> > || =====================
> > || 
> > || mailto "root"
> > || dumpuser "amanda"
> > || netusage  100000 Kbps
> > || maxdumps 1
> > || 
> > || bumpsize 10 Mb
> > || bumpdays 1
> > || bumpmult 4
> > || 
> > || etimeout 1800
> > || 
> > || runtapes 1
> > || reserve 0
> > || 
> > || define tapetype DEC-TZ89 {
> > ||   comment "DEC TZ89 (hardware compression on: Xmn)"
> > ||   length 29414 mbytes
> > ||   filemark 24 kbytes
> > ||   speed 1634 kps
> > || }
> > || 
> > || ======================
> > || 
> > || and the other:
> > || 
> > || ======================
> > || 
> > || org "IS"
> > || 
> > || inparallel 1
> > || 
> > || dumpcycle 8
> > || runspercycle 0
> > || tapecycle 4
> > 
> > Not sure I understand this set of parameters.
> > 
> > runspercycle == 0 causes all level 0 dumps I believe.
> > In that case the "length" of the dumpcycle is immaterial.
> > Each time you run amdump, everything will get full dumps.
> 
> runspercycle = 0 means "same as tapecycle" as mentionend
> in another post here.

That was "me" in the other post.  That is not what I wrote.

> If this is not correct (at least the docu says, it is correct)
> than please tell me. I want amanda to know that there is one
> run every day, despite of manual runs after fixing errors.

The docs (and I) said "same as dumpcycle".  How soon I forget
even what I wrote and read.  Must be a sign of senility.  :(
I was thinking dumpcycle == 0 would cause all level 0 dumps.
With your settings it does suggest one run per day.  Sorry.

I know you are not using tape now, but whenever you run
amcheck (you do don't right?) it will complain tapecycle
less than runspercycle.

> > 
> > Are you saying you plan to run amdump once each 8 days
> > and that you want only level 0 dumps each run?
> > 
> > And further, you will rotate 4 tapes totaling 32 days
> > of backups available?
> 
> I removed all tapes from the config. I just save to the holding disk. I
> can live with the daily "tape error", but I want amanda to do less full
> backups.
> 

The general way to reduce full dump frequency is increase dumpcycle.

If that doesn't help then perhaps other factors are coming into play.
For example, amanda will promote a very active DLE thinking so much
has changed the savings of a level 1 aren't worth it, lets just do
another level 0.

Sometimes an activity on your computer will make it seem like things
have changed when they have not.  I believe gnutar relies heavily
on the ctime parameter of unix files.  Some activities affect ctime
and could make it look like the file changes (ex. moving a file).

Other than that I don't see any reason you are not getting higher
number levels on your holding disk.  I certainly do on a new setup
I'm testing but not providing any tapes.

I don't think this is the problem, but I do not see a "record yes"
setting.  I wonder if that is needed to ensure amanda knows when
the last full dump was done?  It is needed for "dump", but I'm
uncertain about gnutar.


> > 
> > || 
> > || holdingdisk hd1 {
> > ||     comment "main holding disk for IS"
> > ||     directory "/data/amanda/IS"
> > ||     use -1
> > ||     chunksize 2 Gb
> > ||     }
> > || 
> > || 
> > || tapedev "/dev/null"
> > || tapetype DEC-TZ89
> > 
> > Is your DEC tapedrive really attached to the null device?
> > Doesn't work very well that way does it?  :))
> 
> It is disconnected from the server as it produced some driver problems
> recently. So we switched to backup to the holdingsdisk.
> 
> > 
> > 
> > || labelstr "^IS[0-9][0-9]*$"
> > || 
> > || infofile "/var/lib/amanda/IS/curinfo"
> > || logdir   "/var/lib/amanda/IS"
> > || indexdir "/var/lib/amanda/IS/index"
> > || 
> > || includefile "/etc/amanda/amanda-shared.conf"
> > || 
> > || ===================
> > || 
> > || the first file is used by all configs. The second one is one specific
> > || config.
> > || 
> > 
> --


>>> End of included message <<<

-- 
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)