* Abilio Carvalho <abilio.carva...@bbp.ch> [20130430 10:36]:
> sorry, no I hadn't disabled compression. Today I installed mt-st and 
> following the wiki I ran:
> 
> mt -f /dev/sg2 compression 0
> mt -f /dev/sg2 defcompression 0
> 
> but when I now run amtapetype it still tells me:
> 
> Applying heuristic check for compression.
> Wrote random (uncompressible) data at 73376008.2580645 bytes/sec
> Wrote fixed (compressible) data at 101095833.6 bytes/sec    
> Compression: enabled
> 
> how can I make sure hardware compression is off?

Make sure you have a tape loaded. 
Replace the tape drive scsi generic device sgX accordingly.
In case of LTO5 here is what I get:

~# tapeinfo -f /dev/sgX
Product Type: Tape Drive
Vendor ID: 'HP      '
Product ID: 'Ultrium 5-SCSI  '
...
DataCompEnabled: yes
DataCompCapable: yes
DataDeCompEnabled: yes
...

~# mt-st -f /dev/nst0 defcompression 0
~# tapeinfo -f /dev/sgX
...
DataCompEnabled: no
DataCompCapable: yes
DataDeCompEnabled: yes

hth,
jf

> 
> On Apr 17, 2013, at 6:40 PM, Jean-Francois Malouin 
> <jean-francois.malo...@bic.mni.mcgill.ca> wrote:
> 
> > * Abilio Carvalho <abilio.carva...@bbp.ch> [20130417 10:19]:
> >> Well, since noone answered, I had to run amtapetype, so here it is if 
> >> anyone else needs it:
> >> 
> >> define tapetype LTO6 {                                      
> >>    comment "Created by amtapetype; compression enabled"
> >>    length 2442818848 kbytes
> >>    filemark 1806 kbytes
> >>    speed 74006 kps
> >>    blocksize 32 kbytes
> >> }
> > 
> > Hmmm, speed look way off. Native speed is 160MBs for LTO6...
> > Did you disable hardware compression before running amtapetype?
> > Also, you might want to test with larger blocksize, for LTO5 I use 2MB
> > with good results....
> > 
> > hth,
> > jf
> > 
> >> 
> >> 
> >> On Apr 16, 2013, at 3:24 PM, Abilio Carvalho <abilio.carva...@bbp.ch> 
> >> wrote:
> >> 
> >>> Hi, just a quick question, does anyone have a tape type for an LTO6 
> >>> drive? (Ideally an HP Ultrium 6 SCSI)
> >>> 
> >>> Thanks
> >>> Abilio

Reply via email to