[ale] LVM snapshot backups

Jim Kinney jim.kinney at gmail.com
Sat Jan 23 14:09:22 EST 2016


The new data is written to the ORIGINAL lvm. At that point, the old data is
written to the snapshot lvm.  The snapshot lvm is initially of zero real
size, no disk space is used. Only when blocks need to be written to in the
original lvm are blocks written to in the snapshot. Thus drive space used
is double the size of the new data: one write to archive the old data in
the snapshot and a second write to store the new data in the original lvm.
Its a performance hit but an overall time saver. Think of the snapshot as
nothing but hard links to physical extents until an extent changes.
On Jan 23, 2016 12:42 PM, "DJ-Pfulio" <djpfulio at jdpfu.com> wrote:

> Concerning LVM backups.
>
> If all the change data is being written to the snapshot area, how does
> that get
> merged back into the main storage and main DB after we are done with
> backups?
>
>
>
>
> On 01/22/2016 02:40 PM, Jim Kinney wrote:
> > Slide deck is live on ale.org in meeting announcement page.
> > On Fri, 2016-01-22 at 10:54 -0500, DJ-Pfulio wrote:
> >> Are there notes, the presentation, and/or video from last night?
> >> Sounds like the presentation and some notes could really be helpful
> >> based on the
> >> comments.
> >>
> >> Sorry I missed it!
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://mail.ale.org/mailman/listinfo/ale
> See JOBS, ANNOUNCE and SCHOOLS lists at
> http://mail.ale.org/mailman/listinfo
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ale.org/pipermail/ale/attachments/20160123/63b010d3/attachment.html>


More information about the Ale mailing list