Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!wuarchive!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!uunet!mcsun!hp4nl!hpuamsa!frank
From: frank@hpuamsa.UUCP (Frank Slootweg CRC)
Newsgroups: comp.sys.hp
Subject: Re: Question about fbackup
Message-ID: <28510001@hpuamsa.UUCP>
Date: 24 Aug 90 12:29:58 GMT
References: <1204@gufalet.let.rug.nl>
Organization: HP-NL, land of tulips & windmills
Lines: 28

Dear Mr. Roovers,

  A file is defined as active by "fbackup" if the modification time
changes before it finsihes writing the file to tape. I.e., it gets the
modification time, starts writing the file to tape and checks the
modification time again before it finishes.

  Since an active file can not be properly backed_up/restored, "fbackup"
has to retry or skip the active file.

  The number of retries for an active file is defined by "maxretries".
  The maximum amount of media (in bytes) "wasted" by these retries is
limited by "retrylimit".

  Now you might ask *why* is media wasted by a retry? (i.e. Why does
"fbackup" not skip back to the beginning of the "bad" file on tape?)

  I don't know the absolute answer to this but I *think* it is because :

- A file on tape does (most of the time) not start on a physical record
  boundary (i.e. f.e. after a IRG for magtape) and/or
- "fbackup" can have its output piped to another program. It is
  impossible to seek (backward or forward) on a pipe.

  Hope this help. Please call or mail us if you need further help or
information. After all that is what we/software_contracts are for :-)

Frank Slootweg, RCE/Teamleader HP-UX Support, Dutch Customer Response Center.