From: Dave Chinner <david@fromorbit.com>
To: Hans-Peter Jansen <hpj@urpla.net>
Cc: Michael Lueck <mlueck@lueckdatasystems.com>, linux-xfs@vger.kernel.org
Subject: Re: Selectively backup based on specific date
Date: Fri, 24 Mar 2017 08:38:31 +1100 [thread overview]
Message-ID: <20170323213831.GF17542@dastard> (raw)
In-Reply-To: <1904662.EPkO7u6dZY@xrated>
On Thu, Mar 23, 2017 at 03:11:07PM +0100, Hans-Peter Jansen wrote:
> Hi Michael,
>
> On Donnerstag, 23. März 2017 09:21:56 Michael Lueck wrote:
> > Greetings,
> >
> > I am looking for the correct xfsdump syntax to backup dirs/files newer than
> > a certain date.
>
> xfsdump is a tool in the tradition of unix filesystem dump utilities.
> Those always dump a full filesystem, and consequently, fully restore
> filesystems only.
Not true. xfsdump/restore can backup directory subtrees just fine.
See, for example, the '-s pathname' option....
> > I have not found mention of such an option / capability within xfsdump.
> > Hopefully there is a creative way to do the selective backup.
xfsdump does "backup files changed since" operations via incremental
backups. i.e. if you want to backup only things changed since $DATE,
you first need to have run a dump that on $DATE, and you then do an
incremental dump which will contain only the changed files/dirs
since $DATE...
IOWs, xfsdump is designed to be used as part of a well-defined,
managed backup and disaster recovery strategy - it's not an ad-hoc
"copy random specification of files" tool.
> Using a linux derivate, there are plenty of more or less creative options to
> selectively backup/restore arbitrary file systems, but this is rather OT for
> the ML.
The downside of these generic utilities is they are much slower than
xfsdump for large scale backup operations, they trash the
inode/dentry/page caches and they don't really support concurrent
operations needed to scale effectively. i.e. if you need to dump
throughput of multiple GB/s, process tens/hundreds of millions of
files and/or drive multiple tape drives concurrently from a single
filesystem, then xfsdump is a much better option...
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
prev parent reply other threads:[~2017-03-23 21:38 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-03-23 13:21 Selectively backup based on specific date Michael Lueck
2017-03-23 14:11 ` Hans-Peter Jansen
2017-03-23 14:31 ` Michael Lueck
2017-03-23 21:38 ` Dave Chinner [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20170323213831.GF17542@dastard \
--to=david@fromorbit.com \
--cc=hpj@urpla.net \
--cc=linux-xfs@vger.kernel.org \
--cc=mlueck@lueckdatasystems.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.