Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.misc > #4984

Re: Online backups

From Paul Sture <nospam@sture.ch>
Newsgroups comp.misc
Subject Re: Online backups
Date 2014-09-19 13:14 +0200
Organization PostgreSQL and SQLite
Message-ID <jpfteb-ifj2.ln1@news.chingola.ch> (permalink)
References <4jdi1ap1m057d9ociq7a9nusdq5gh0g07l@4ax.com> <6858#39341@usenet.squte.com> <lveh6u$q9n$1@dont-email.me> <20140918133113.4e623011.rsw@therandymon.com> <lvfpbe$73i$1@dont-email.me>

Show all headers | View raw


On 2014-09-18, Rich <rich@example.invalid> wrote:
> RS Wood <rsw@therandymon.com> wrote:
>> On Thu, 18 Sep 2014 11:55:10 +0000 (UTC)
>> Rich <rich@example.invalid> wrote:
>
>> > RS Wood <rsw@therandymon.com> wrote:
>> > > But my pictures and movies I burn to DVDs or CDs every few months,
>> > > each time starting from the present and going back as far as I can
>> > > while still fitting on the disk. That way there's always some overlap
>> > > and multiple disks have part of that overlap, affording me some
>> > > redundancy.
>> > 
>> > For additional redundancy on the DVD's/CD's, you might consider
>> > Dvdisaster (http://dvdisaster.net/en/index.html) as well as your
>> > current "overlapping" process.
>
>> Yes, Scott mentioned it above and it caught my attention.  I'm going to
>> have to take some time to figure it out, but I'm already interested.
>> Thanks for providing a link.
>
> Your welcome.  It uses part of the storage space on the disk
> (configurable) to store reed-solomon error correcting data such that a
> percentage of the disk can be unreadable and yet you can still recover
> all the data you wrote to the disk.

Ah that's interesting.  For many years I was using a backup solution with
built in CRC (and used by default) and having seen how effective that was
with tapes throwing read errors I rather miss it.  Once tape drives which
supported their own CRC came out various folks started to recommend
switching off the CRC in the backup utility but the latter represented the
full path from disk to tape so I stuck with it (and had systems powerful
enough that it wasn't a performance issue).

> It also handles the "recovery" process should you need to recover a
> disk.

Good to hear.  A recovery process that is tricky to use gets too close to
performing write-only backups.

> It is to help against the "disk-rot" effect that plagues so many
> writable optical media.  Over time, the dyes "rot" away sufficiently
> that you can't read everything from the disk anymore.  Thankfully, for
> disks where this does happen, it is usually a slow process and happens
> for only small parts at a time.  So provided you test read the disks
> occasionally, you can recover from rot before all it lost.

We were discussing deduplication of backups a couple of weeks ago and I
came to the conclusion that redundancy isn't such a bad thing.
Bit rot on the backups could mean that a vital block present across many
backups and deduplicated to one copy only could make life difficult.

> Or at least that is the intent.  If you never ever test the old disks,
> you'll never know if they are readable or not, until you need them, and
> then, if the answer is not, you end up stuck.
>
>> Hell, at less than $1 per disk these days it's not even that much
>> money.

I was sure they cost substantially cost more here, at least the last time
I looked, but I can get stacks of 50 or 100 at that price (just not any
more in high street outlets).

> It is also just about your only option beyond building a raid array
> backup server to do backups of todays datasets when one is usually on
> dialup (you've mentioned your dialup situation before).  Online backup
> is just not an option at dialup speeds.

It often isn't a viable option at the lower broadband speeds either.

-- 
Ernesto "Che" Guevera, commonly known as el Che, was an Argentinian
t-shirt tycoon.                                  -- TL;DR Wikipedia

Back to comp.misc | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread


Thread

Online backups Chris J Dixon <chris@cdixon.me.uk> - 2014-09-17 08:24 +0100
  Re: Online backups scott@alfter.diespammersdie.us (Scott Alfter) - 2014-09-17 15:44 +0000
  Re: Online backups "RS Wood" <rsw@therandymon.com> - 2014-09-18 06:36 -0500
    Re: Online backups Rich <rich@example.invalid> - 2014-09-18 11:55 +0000
      Re: Online backups RS Wood <rsw@therandymon.com> - 2014-09-18 13:31 +0000
        Re: Online backups Rich <rich@example.invalid> - 2014-09-18 23:20 +0000
          Re: Online backups Will Honea <whonea@whonea.net> - 2014-09-18 18:37 -0600
          Re: Online backups Paul Sture <nospam@sture.ch> - 2014-09-19 13:14 +0200
            Re: Online backups Will Honea <whonea@whonea.net> - 2014-09-19 12:51 -0600
          Re: Online backups scott@alfter.diespammersdie.us (Scott Alfter) - 2014-09-19 15:43 +0000
            Re: Online backups Rich <rich@example.invalid> - 2014-09-19 18:50 +0000
              Re: Online backups scott@alfter.diespammersdie.us (Scott Alfter) - 2014-09-19 21:43 +0000
      Re: Online backups scott@alfter.diespammersdie.us (Scott Alfter) - 2014-09-18 17:06 +0000
    Re: Online backups scott@alfter.diespammersdie.us (Scott Alfter) - 2014-09-18 16:27 +0000
  Re: Online backups Roger Blake <rogblake@iname.invalid> - 2014-09-19 16:31 +0000
    Re: Online backups Chris J Dixon <chris@cdixon.me.uk> - 2014-09-20 08:03 +0100
      Re: Online backups Roger Blake <rogblake@iname.invalid> - 2014-09-22 16:34 +0000
        Re: Online backups Chris J Dixon <chris@cdixon.me.uk> - 2014-09-22 19:28 +0100

csiph-web