Re: Split up dumps for DVD-RAM

On 2006-10-04 13:45, mario wrote:


Here is my config: http://pastebin.ca/190833


You still have runtapes 1, and some more weird/incomplete
parameters:


 runtapes 1
 [...]
 tapetype HARD-DISK
 [...]
 define dumptype user-tar-span {
  root-tar
  tape_splitsize 3 Gb
  comment "tape-spanning user partitions dumped with tar"
  priority medium
 }
 [...]
 define tapetype DVD47G {
  length 4350 MB
  filemark 4KB
 }
 define tapetype HARD-DISK {
  comment "Dump onto hard disk"
  length 3072 mbytes  # specified in mbytes to get the exact size of GB
 }



First, are you sure you want "tapetype HARD-DISK"and not
"tapetype DVD47G"? Why wasting 1.5 Gbyte of each DVD?


FAILURE AND STRANGE DUMP SUMMARY:
  server  md1  lev 0  FAILED [dump larger than available tape space,
61240550 KB, but cannot incremental dump new disk]


OK, indeed.  Amanda needs to write 6 Gbyte but there is only
one vtape of 3Gbyte available.

Specifying "runtapes 3" will solve that problem.
Then amanda will write 3 vtapes, each of max size 3Gbyte, which
you can burn to DVD.

Note also that "tape_splitsize" is the size of EACH chunk that will
be written to tape.  Because your vtape is only 3 GByte
and one chunk is also 3 GByte, you're already missing space for
the 32Kbyte label at the start of each tape.
(Currently, Amanda write to tape until it hits into EOT, even so
for a virtual tape.  The last, incomplete, chunk will not be erased
from that vtape.  And that same chunk will be started over again
on the next tape.  When you specify a large size for the splits,
that results in a lot of wasted space (and time to write those).

I would "tape_splitsize 512m" instead.
And I would add:  split_diskbuffer "/path/to/scratch/with512M/free".
(see below)

Also do something about these errors:

NOTES:
  planner: tapecycle (5) <= runspercycle (20)


runspercycle should be larger than tapecycle
In your case, when you burned a DVD from vtape on disk,
you still have the backups, but chg-disk insists on a changer
that has "tapecycle" number of slots.  That's why I would
use chg-multi instead, so that the definitino of tapecycle
is again independent of the number of slots in the changer.

  driver: WARNING: /backup/amanda: not 102400 KB free.


Your holding disk is too small:  100Mbyte only.
Resulting in the trouble below

  planner: disk server:md1, full dump (61240550KB) will be larger than
available tape space


OK, so increase runtapes.

  taper: no split_diskbuffer specified: using fallback split size of
10240kb to buffer (null) in-memory


Amanda has to buffer each chunk before writing it to tape.
Normally, she would dump to the holdingdisk before writing to tape,
that holdingdisk file being the buffer.
But because you specified a holdingdisk that is too small,
she has to fall back to other buffering methods.
Second best is to buffer one "tape-chunk" (with size tape_splitsize)
to a temporary file in a directory specfied by "split_diskbuffer".
But you did not specify that either.
As a last resort Amanda needs to buffer the data in memory, which
by default has a rather conservative size of 10Mbyte.

That last method also results in a lot of small splitted files on tape,
not what you intended either.


--
Paul Bijnens, xplanation Technology Services        Tel  +32 16 397.511
Technologielaan 21 bus 2, B-3001 Leuven, BELGIUM    Fax  +32 16 397.512
http://www.xplanation.com/          email:  Paul.Bijnens AT xplanation DOT com
***********************************************************************
* I think I've got the hang of it now:  exit, ^D, ^C, ^\, ^Z, ^Q, ^^, *
* F6, quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, *
* stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt,  abort,  hangup, *
* PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e,  kill -1 $$,  shutdown, *
* init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... *
* ...  "Are you sure?"  ...   YES   ...   Phew ...   I'm out          *
***********************************************************************