Re: block chg tracking
From: Jeremy Schneider <jeremy.schneider_at_ardentperf.com>
Date: Fri, 13 Feb 2015 14:25:07 -0600
Message-ID: <CA+fnDAZevv5VPsknGNVDip9KQnH_c0KXM9FSuMM0eZAsduNHVg_at_mail.gmail.com>
On Thu, Feb 12, 2015 at 1:18 PM, Mladen Gogala <dmarc-noreply_at_freelists.org> wrote:
> Jeremy, you have to restore one full backup. That's given. Restoring 3 or 4
> additional incremental backups will take significant amount of time. And
> that time can be eliminated using de-duplication technology, because you get
> the best of both worlds: you are only backing up changes and you have a full
> backup.
Date: Fri, 13 Feb 2015 14:25:07 -0600
Message-ID: <CA+fnDAZevv5VPsknGNVDip9KQnH_c0KXM9FSuMM0eZAsduNHVg_at_mail.gmail.com>
On Thu, Feb 12, 2015 at 1:18 PM, Mladen Gogala <dmarc-noreply_at_freelists.org> wrote:
> Jeremy, you have to restore one full backup. That's given. Restoring 3 or 4
> additional incremental backups will take significant amount of time. And
> that time can be eliminated using de-duplication technology, because you get
> the best of both worlds: you are only backing up changes and you have a full
> backup.
My understanding of Commvault Dedupe is that you have to run a full backup (read the whole disk and transfer the whole thing compressed over the network), but commvault will only *store* the changes, using an index to identify duplicate blocks. Is this correct?
The most interesting solutions I've seen allow you to only run incrementals [forever] or only ship logs [forever] and the appliance synthetically generates fulls which can be restored in one shot.
-- http://about.me/jeremy_schneider -- http://www.freelists.org/webpage/oracle-lReceived on Fri Feb 13 2015 - 21:25:07 CET