/dev/rmt/0cbn or /dev/rmt/0mbn?

/dev/rmt/0cbn or /dev/rmt/0mbn?

Post by Mathew Kirsc » Wed, 25 Jul 2001 01:23:10



I'm gonna try it for myself anyway, but in Networker 5.0.1, the device listed
for my tape backup device is /dev/rmt/0mbn.

The drive is a Sun DLT 4700, with a native capacity of 20GB per tape, 40GB
compressed. I always seem to get between 19GB and 21GB on a tape. The problem
is, most of the data being backed up is highly-compressible ASCII source code!
Obviously, there is no compression going on here.

If I use /dev/rmt/0cbn instead of 0mbn, shouldn't I be able to get more than
20GB of ASCII data on a tape?

 
 
 

/dev/rmt/0cbn or /dev/rmt/0mbn?

Post by Darren Dunha » Wed, 25 Jul 2001 05:46:29



> I'm gonna try it for myself anyway, but in Networker 5.0.1, the device listed
> for my tape backup device is /dev/rmt/0mbn.
> The drive is a Sun DLT 4700, with a native capacity of 20GB per tape, 40GB
> compressed. I always seem to get between 19GB and 21GB on a tape. The problem
> is, most of the data being backed up is highly-compressible ASCII source code!
> Obviously, there is no compression going on here.

Is there any chance you're using client side compression?  Are you using
UNIX with compression directives?

Quote:> If I use /dev/rmt/0cbn instead of 0mbn, shouldn't I be able to get more than
> 20GB of ASCII data on a tape?

Yes. (assuming you have a decent st.conf file, which depends on your
version of the OS, patches, and changes you've made to it).  Are you
using CompacTape IV cartridges?

--

Unix System Administrator                    Taos - The SysAdmin Company
Got some Dr Pepper?                           San Francisco, CA bay area
          < How are you gentlemen!! Take off every '.SIG'!! >

 
 
 

/dev/rmt/0cbn or /dev/rmt/0mbn?

Post by Joerg Schilli » Wed, 25 Jul 2001 08:01:48





>#   Is there any chance you're using client side compression?  Are you using
>#   UNIX with compression directives?
>#  
>#  > If I use /dev/rmt/0cbn instead of 0mbn, shouldn't I be able to get more than
>#  > 20GB of ASCII data on a tape?
>#  
>#   Yes. (assuming you have a decent st.conf file, which depends on your
>#   version of the OS, patches, and changes you've made to it).  Are you
>#   using CompacTape IV cartridges?

>And, out of curiosity, does the DLT have
>an LED for compression-is-selected?

I don't know the old ones, but the DLT7000 has many leds, one for compression.

Make sure that the drive streams if possible or it will wear out early...

--



URL:  http://www.fokus.gmd.de/usr/schilling    ftp://ftp.fokus.gmd.de/pub/unix

 
 
 

/dev/rmt/0cbn or /dev/rmt/0mbn?

Post by Mathew Kirsc » Thu, 26 Jul 2001 00:26:58


Looking at the savegroup completion reports, I can see that the amount of data
Networker backs up is exactly half what "df" reports, and all the data is
there on the tape.

When Networker says it's getting 21GB of data on a tape, it's 21GB of
COMPRESSED data...

 
 
 

/dev/rmt/0cbn or /dev/rmt/0mbn?

Post by Darren Dunha » Thu, 26 Jul 2001 02:21:13



> Looking at the savegroup completion reports, I can see that the amount of data
> Networker backs up is exactly half what "df" reports, and all the data is
> there on the tape.
> When Networker says it's getting 21GB of data on a tape, it's 21GB of
> COMPRESSED data...

That's only true if you use compression directives (software
compression) on the client.  If you do, then you are doing the
compression work on the client.  Don't try to use hardware compression
on that stream.  It's not likely to compress any further.

You should probably pick one side or the other.

--

Unix System Administrator                    Taos - The SysAdmin Company
Got some Dr Pepper?                           San Francisco, CA bay area
          < How are you gentlemen!! Take off every '.SIG'!! >

 
 
 

/dev/rmt/0cbn or /dev/rmt/0mbn?

Post by Mathew Kirsc » Thu, 26 Jul 2001 22:06:21



> That's only true if you use compression directives (software
> compression) on the client.  If you do, then you are doing the
> compression work on the client.  Don't try to use hardware compression
> on that stream.  It's not likely to compress any further.

Yep, I'm using compression directives. Compression is done on the backup
client (not the server) right? I figure it's faster to have many fast machines
doing compression in parallel than it would be to make an older tape drive
like that DLT4700 do it all by itself.
 
 
 

/dev/rmt/0cbn or /dev/rmt/0mbn?

Post by Darren Dunha » Sat, 04 Aug 2001 10:13:41




>> That's only true if you use compression directives (software
>> compression) on the client.  If you do, then you are doing the
>> compression work on the client.  Don't try to use hardware compression
>> on that stream.  It's not likely to compress any further.
> Yep, I'm using compression directives. Compression is done on the backup
> client (not the server) right? I figure it's faster to have many fast machines
> doing compression in parallel than it would be to make an older tape drive
> like that DLT4700 do it all by itself.

Maybe.  Remember the DLT has the compression routines in silicon.  It
can probably compress the data faster than it can write it on the tape.  

In the old days, I'd tell you you were crazy to try it.  Nowadays, you
can probably do it and win.  The clients are getting faster, but the
network isn't getting bigger as fast.  You're saving some network
bandwidth too, but then again, a DLT 4000 isn't that fast either..

If you do that, then you probably want to double check the st.conf that
you're using.  Generally the 'h' device will be set for the highest
density level, but non-compressing.  That may be the best bet.  On a
4000, that may be the same as the 'm' device though...

--

Unix System Administrator                    Taos - The SysAdmin Company
Got some Dr Pepper?                           San Francisco, CA bay area
          < How are you gentlemen!! Take off every '.SIG'!! >

 
 
 

1. Cannot open /dev/rmt/0un (or any other /dev/rmt/??n)

(I see this group is much more active)

I've moved an L9 autoloader from a Solaris 9 E250 to a Solaris 9 V120.

I've duplicated the info from st.conf

tape-config-list=
"QUANTUM DLT8000","Quantum DLT8000","DLT8k-data";
DLT8k-data = 1,0x38,0,0x10639,4,0x84,0x85,0x88,0x89,3;

and sgen.conf

device-type-config-list="changer", "sequential"
.....
name="sgen" class="scsi" target=0 lun=0;
name="sgen" class="scsi" target=1 lun=0;
name="sgen" class="scsi" target=2 lun=0;
name="sgen" class="scsi" target=3 lun=0;
name="sgen" class="scsi" target=4 lun=0;
name="sgen" class="scsi" target=5 lun=0;
name="sgen" class="scsi" target=6 lun=0;
name="sgen" class="scsi" target=7 lun=0;
name="sgen" class="scsi" target=8 lun=0;

# ls -la /dev/scsi
total 14
drwxr-xr-x   4 root     other        512 May 16 08:31 .
drwxr-xr-x  15 root     sys         4096 May 17 07:26 ..
drwxr-xr-x   2 root     other        512 May 16 08:31 changer
drwxr-xr-x   2 root     other        512 May 16 08:31 sequential

My problem is every time I do a backup I get the following:

  DUMP: Date of this level 0 dump: Wed May 17 10:44:21 2006
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c1t0d0s0 (west:/) to /dev/rmt/0un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Writing 63 Kilobyte records
  DUMP: Estimated 269898 blocks (131.79MB).
  DUMP: NEEDS ATTENTION: Cannot open `west:/dev/rmt/0un'.  Do you want
to retry the open?: ("yes" or "no") no
  DUMP: The ENTIRE dump is aborted.
  DUMP: Date of this level 0 dump: Wed May 17 10:44:24 2006
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c1t0d0s4 (west:/var) to /dev/rmt/0un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Writing 63 Kilobyte records
  DUMP: Estimated 544350 blocks (265.80MB).
  DUMP: NEEDS ATTENTION: Cannot open `west:/dev/rmt/0un'.  Do you want
to retry the open?: ("yes" or "no")

When the L9 was connected to the E250, I had no problems.  I tried
every /dev/rmt/ that ended with an "n" and all return the same error.

I think I am overlooking something obvious but so far no luck on a
solution.

Thanks!

2. Is this possible?

3. tape device bulebud01 / 57 <> mt status /dev/rmt/3n /dev/rmt/0n: Device busysy?

4. possible issues on SMP machine for Netfilter based kernel module

5. tape drive /dev/rst11 instead of /dev/rmt/0

6. A Y2K consideration

7. What is the difference between /dev/rmt/0 and /dev/rst5?

8. ld.so ???

9. Linux rmt incompatible with Solaris rmt

10. Empty /dev/rmt on Solaris 2.1 / x86 HELP

11. Slow writes to /dev/rmt/0c

12. rsh remote dd if=/dev/rmt/0 bs=1024k | tar -xvf -

13. dev/rmt/0