Problems with dump/capacity/DLT IV 40-80GB Tape cartridges

Andreas Totlis usenet at
Mon Apr 7 01:03:58 PDT 2003

Hello Nathan,

I'm always fight with my SLR50 Tape Drive :-)

* Nathan Vidican <nathan at> [030406 16:49]:
> sa0 at ahc1 bus 0 target 1 lun 0 
> sa0: <BNCHMARK DLT1 391B> Removable Sequential Access SCSI-2 device 
> sa0: 20.000MB/s transfers (10.000MHz, offset 8, 16bit) 
> Output of 'df' returns the following: 
> Filesystem    1K-blocks    Used    Avail Capacity  Mounted on 
> /dev/aacd0s1a   2064302   64488  1834670     3%    / 
> /dev/aacd2s1e  70558134 4227862 60685622     7%    /server 
> /dev/aacd0s1f   6132382 1363156  4278636    24%    /usr 
> /dev/aacd0s1e   8257742  124818  7472306     2%    /var 
> procfs                4       4        0   100%    /proc 
> The Problem: 
> ~~~~~~~~~~~~ 
> When using 'dump -0au -f /dev/sa0 /server' I cannot fit the contents of 
> /server (approx 4.2gigs of data) onto a single tape volume. The tape 
> inserted is a brand new HP C5141F, (which is a 40/80GB DLT IV cartidge). 
> Dump prompts for a new tape to be inserted to complete the operation. I 
> assumed this had to do with the '-a' part in the dump command; assuming that 
> dump is therefore not (properly) auto-detecting the EOT (end of tape), or 
> using the wrong recording density and reaching the end of the tape all too 
> quickly. Using 'restore -i' I can indeed read the data written to the tape, 
> so it is writting the data... just not properly. 

> The Question: 
> ~~~~~~~~~~~~~ 
> Is anyone out there using similar hardware, or know of the proper arguments 
> to pass to dump to make full use of the capacity of these DLT IV cartridges? 
> I have checked HP's website, and these particular cartidges list a length of 
> 557Meters, and a density of 46.8K/80K/96K dependant upon 40GB/70GB/80GB 
> compression schemes. How do I formulate the dump command to properly make 
> use of these tapes.... any suggestions? 

Try to set the blocksize by your Drive(e.g.: "mt -f /dev/... blocksize
64" - see man mt). Set the blocksize in your dump command.

My actually dump call: 
/sbin/dump 0aubf 512 - /mail | /usr/local/bin/team 1m 3 >/dev/nrsa0

Anymore you should use a buffer (e.g. team) to ensure that the
Drive streams continous.


Andreas -

More information about the freebsd-questions mailing list