[Dump-users] blocks vs records: comments and questions

[Abner G. <67...@gm...>]

I am a relatively recent user of dump and am looking for some confirmation
that I understand this application correctly as some of it seems a bit
unusual.

I have spent some time studying on blocks vs records with regards to such
programs as tar, dd, and more recently dump. There are several features of
dump that are a bit confusing in this regard. I should say that I am using
version 0.4b44 under Debian linux.

Here is a portion of the dump output:

  DUMP: Label: xyz
  DUMP: Writing 512 Kilobyte records
  DUMP: mapping (Pass I) [regular files]
  DUMP: mapping (Pass II) [directories]
  DUMP: estimated 91547938 blocks.
  DUMP: Volume 1 started with block 1 at: Sun May  1 12:29:52 2016
  DUMP: 0.00% done, finished in 0:00
  DUMP: dumping (Pass III) [directories]
  DUMP: dumping (Pass IV) [regular files]
  DUMP: 21.45% done at 65448 kB/s, finished in 0:18
  DUMP: 42.84% done at 65372 kB/s, finished in 0:13
  DUMP: 69.20% done at 70388 kB/s, finished in 0:06
  DUMP: Closing /dev/nst0
  DUMP: Volume 1 completed at: Sun May  1 12:49:20 2016
  DUMP: Volume 1 91489792 blocks (89345.50MB)
  DUMP: Volume 1 took 0:19:28
  DUMP: Volume 1 transfer rate: 78330 kB/s


1. Dump announces that it is using 512 K record size but then gives
completion information in number of 1 K blocks. Dump writes 512 one
kilobyte blocks to each record. Why not just announce result as "91489792
kB" rather than "91489792 blocks" or better yet provide number of records
written to tape? Was this of some historic significance?

2. Is it possible to alter the block size to something larger than 1 K?
Does not appear to be the case or at least not documented in man page.

3. Option -b is referred to in man page as "blocksize" but it really sets
the record size. This does however directly relate to bs= option of dd
command.

4. Option -B is referred to as "records" but it is really asking to specify
the number of 1 K blocks that can fit on the tape.

5. There might be value in recording the number of records in the dump for
verification or data recovery purposes but dump does not provide this info
directly. The size of the dump is output in kB and MB. From these numbers
the number of records can be calculated assuming there are no extra records
created for headers or at end of the dump file. Does anyone bother tracking
number of records per dump? I usually just record the MB.

Lastly, I want to say thank you to the authors, maintainers, and other
contributors as I am greatly appreciative for this useful application.

Regards,
Abner