Bacula-users

Re: [Bacula-users] Backup of many many files

2008-09-04 15:55:31
Subject: Re: [Bacula-users] Backup of many many files
From: Ryan Novosielski <novosirj AT umdnj DOT edu>
To: bacula-users AT lists.sourceforge DOT net
Date: Thu, 04 Sep 2008 15:54:58 -0400
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Arno Lehmann wrote:
> Hi,
> 
> 04.09.2008 14:00, T. Horsnell wrote:
>> FMI, will someone please tell me how shoeshining manifests itself?
>> Does it mean that the tape actually reverses periodically?
> 
> Yes.
> 
>> e.g. to backspace over a filemark in order to overwrite it, or does it 
>> mean that the tape just stop-starts rather than continuously streaming. 
>> If its just stop-start, why does it result in more tape wear (as opposed 
>> to tapedrive wear) than streaming?
> 
> Because a tape can not be stopped and started in an instant - the tape 
> has mass, it moves with high speed, so even if you had a drive 
> mechanism that would be capable of starting and stopping 
> (near-)instantaneously, the tape would break.
> 
> What happens is this:
> - Tape is driven at writing speed.
> - Drive runs out of data and, after the last byte in its buffer is 
> written, it stops the tape. Some tape passes until the tape comes to rest.
> - Drive gets more data
> - Drive rewinds the tape until safely before the current end-of-data 
> on tape
> - Drive starts moving the tape forward
> - Once read speed is reached, the drive syncronizes to the tape tracks 
> and reads until the current end-of-data is reached.
> - In this instant it switches to write mode and writes the current data.
> Sometimes the first attempt to find the writing position will not 
> work, so this process can iterate a few times.
> The result is that the tape passes the read/write heads a - sometimes 
> huge - number of times. Each pass rubs some matter from the tape and 
> the heads.
> Furthermore, the strain on the tape itself is highest when accelerated 
> and stopped.
> 
> Todays tape technology, which writes hundreds of MB in seconds, really 
> suffers if you feed bursts of data to it - a GB of data, fed in small 
> bursts, can result in dozens of start-stop-cycles.
> 
> Another effect is that, when the write position is found, a small 
> length of tape can pass unused before actual data is written again, so 
> you lose tape capacity. Similarly, a file mark on tape can use a 
> considerable length of tape - both effects decrease the useable tape 
> capacity.
> 
> So you really want to avoid this shoeshining and lots of file marks.

Thank you for this excellent description. I learned a little myself.

- --
 ---- _  _ _  _ ___  _  _  _
 |Y#| |  | |\/| |  \ |\ |  | |Ryan Novosielski - Systems Programmer II
 |$&| |__| |  | |__/ | \| _| |novosirj AT umdnj DOT edu - 973/972.0922 (2-0922)
 \__/ Univ. of Med. and Dent.|IST/AST - NJMS Medical Science Bldg - C630
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFIwD0Smb+gadEcsb4RAjS0AKCnnlg6XYAAwf7PO2USlIkP2rqJUACfSRu4
IiX74nx1oHbkwtKt1UBuClU=
=5HOz
-----END PGP SIGNATURE-----

Attachment: novosirj.vcf
Description: Vcard

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users