Veritas-bu

Re: [Veritas-bu] NDMP and Millions of Files

2010-06-21 08:54:14
Subject: Re: [Veritas-bu] NDMP and Millions of Files
From: Ed Wilts <ewilts AT ewilts DOT org>
To: Bahadir Kiziltan <bahadir.kiziltan AT gmail DOT com>
Date: Mon, 21 Jun 2010 07:48:12 -0500
On Mon, Jun 21, 2010 at 2:19 AM, Bahadir Kiziltan <bahadir.kiziltan AT gmail DOT com> wrote:
Try NetApp NDMP streamer, which allows you to leverage the deduplication in NDMP backups.

You need at least PureDisk 6.6.x + NBU 6.5.4 with EEB.

The PureDisk requirement is the killer. I've got several 20TB applications with hundreds of millions of files.  Purchasing PureDisk licenses for that much data is prohibitively expensive.

As long as Symantec insists on a per-TB license for PureDisk, we will continue to make as little use of it as we can.

   .../Ed

Ed Wilts, RHCE, BCFP, BCSD, SCSP, SCSE
ewilts AT ewilts DOT org
 


On Thu, Jun 17, 2010 at 5:58 PM, <ccosta.ccc AT gmail DOT com> wrote:
Good news is that with 7.01 coming out in a month or so will allow NBU to multi-stream NDMP data to a single tape drive.

This will/may alleviate some of the performance issues many of you experience each day. However I am not sure of any limitations of this feature patch may have.

Chris

Sent from my Verizon Wireless BlackBerry

-----Original Message-----
From: "Ambrose, Monte" <mambrose AT qualcomm DOT com>
Date: Thu, 17 Jun 2010 14:38:41
To: Jonathan Dyck<jdyck AT bank-banque-canada DOT ca>; Jeff Cleverley<jeff.cleverley AT avagotech DOT com>; Rusty.Major AT sungard DOT com<Rusty.Major AT sungard DOT com>
Cc: veritas-bu AT mailman.eng.auburn DOT edu<veritas-bu AT mailman.eng.auburn DOT edu>
Subject: Re: [Veritas-bu] NDMP and Millions of Files

You could also use NetBackups Snap Mirror to tape.  You would need a NetApp snap mirror license.

The Pros
It uses snapmirror and sends the data off to tape.  It is a RAW volume backup and is extremely fast - 10X in many cases.
It can be fully configured in NetBackup

The Cons
It backs up the entire volume - so if you have a 1TB volume and only 400GB are used it will backup 1tb
You cant restore a single file or dir.  You have to restore the entire thing.
You cant mix with incremental backups.

Monte

-----Original Message-----
From: veritas-bu-bounces AT mailman.eng.auburn DOT edu [mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of Jonathan Dyck
Sent: Tuesday, June 15, 2010 7:21 AM
To: Jeff Cleverley; Rusty.Major AT sungard DOT com
Cc: veritas-bu AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu] NDMP and Millions of Files

We do our NetApp backups (old GF940 metro cluster) with a combination of snapshots (which are available at both sites) and NDMP here. Our NDMP obviously holds all the long retention data.

The data size isn't huge compared to some (6.5TB,  ~40M files), but we've had to resort to multiple policies and multi-streaming to back it up in a reasonable amount of time (less than 60 hours for a full on the wknd).  The way it works is:

PolicyA (vol1): explicitly lists 37 different paths for the backup selection list,  we've empirically determined these are the "smaller" folders
PolicyB (vol1-long): explicitly lists 16 paths for the backup selection list,  we've determined these are the "large" folders
PolicyC (vol1-Catch-missed-directory): we've mounted the root of vol1 on a Linux host,  and we back it up via NFS, excluding the 37+16 paths defined above.  If this policy's full backup every gets too large (over 10GB or so),  we review the contents and add new paths to PolicyA or PolicyB as necessary.  This is necessary because you can't specify wildcards on NDMP backups (discussed in this forum several times I believe).

We repeat the above process for vol2.

The above backup data sits on a deduped VTL for 2 months, and then the data that is held longer than that is duplicated to tape for long-term storage and expired off the VTL.  As we run 5 streams concurrently, the throughput is decent, but we peg out the CPU on the NetApp frequently during the backup window,  which is a concern.

HTH...
Jon



-----Original Message-----
From: veritas-bu-bounces AT mailman.eng.auburn DOT edu [mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of Jeff Cleverley
Sent: June 14, 2010 5:57 PM
To: Rusty.Major AT sungard DOT com
Cc: veritas-bu AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu] NDMP and Millions of Files

Rusty,

If you have a way to use Snapvault to backup to another location, I
would use it.  We have a number of file systems like what you have.  I
tried NDMP over TCP and NFS backups using dedicated snapshots mounted
on a client.  Both used a dedicated 10G network.  We basically overran
our 6030 filer.  We could have jumped through a lot of hoops and split
backups over multiple weekends, etc, but we decided it wasn't worth
it.  I haven't tried Flash Backup for a while but it didn't really buy
us much on what we tried to do with it.  It may work better now.

We backup everything (~200 TB) to NearStores in another building.  We
use SnapVault instead of SnapMirror.  We can still revert our
destination volumes to primary r/w file systems if we need to.  We
don't have the requirement to send tapes off site.  If you do, you
could still make the tapes from your secondary filer.

Using NFS to tape will give you check points and you can run multiple
streams to each tape drive.  If you are not hitting the filer
throughput limits this may work for you.

Jeff

On Mon, Jun 14, 2010 at 3:24 PM,  <Rusty.Major AT sungard DOT com> wrote:
>
> We have a NetApp filer that has a few TB of data made up largely of millions
> of small files (about 30 million or so) and we are using several NDMP
> policies to back up this data. The two main problems are length of time it
> takes to backup (we usually have 2-3 backups running all day every day) and
> when there is a maintenance or other event in the NBU domain, we have to
> kill the job, resulting in having to start all over (no checkpoints).
>
> For those of you who have faced a similar situation, how are you backing up
> this data?
>
> Current thoughts are moving away from NDMP and going with just snapshots and
> then getting the snap offsite either by backing it up or replicating it.
> We've also thought about backing it up via NFS, but that will probably be
> slower, though we would get checkpoints.
>
> I appreciate any other suggestions anyone has.
>
> Rusty Major, MCSE, BCFP, VCS ▪ Sr. Storage Engineer ▪ SunGard Availability
> Services ▪ 757 N. Eldridge Suite 200, Houston TX 77079 ▪ 281-584-4693
> Keeping People and Information Connected® ▪ http://availability.sungard.com/
> P Think before you print
> CONFIDENTIALITY:  This e-mail (including any attachments) may contain
> confidential, proprietary and privileged information, and unauthorized
> disclosure or use is prohibited.  If you received this e-mail in error,
> please notify the sender and delete this e-mail from your system.
> _______________________________________________
> Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
>
>



--
Jeff Cleverley
Unix Systems Administrator
4380 Ziegler Road
Fort Collins, Colorado 80525
970-288-4611
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
====================================================================================

La version française suit le texte anglais.

------------------------------------------------------------------------------------

This email may contain privileged and/or confidential information, and the Bank of
Canada does not waive any related rights. Any distribution, use, or copying of this
email or the information it contains by other than the intended recipient is
unauthorized. If you received this email in error please delete it immediately from
your system and notify the sender promptly by email that you have done so.

------------------------------------------------------------------------------------

Le présent courriel peut contenir de l'information privilégiée ou confidentielle.
La Banque du Canada ne renonce pas aux droits qui s'y rapportent. Toute diffusion,
utilisation ou copie de ce courriel ou des renseignements qu'il contient par une
personne autre que le ou les destinataires désignés est interdite. Si vous recevez
ce courriel par erreur, veuillez le supprimer immédiatement et envoyer sans délai à
l'expéditeur un message électronique pour l'aviser que vous avez éliminé de votre
ordinateur toute copie du courriel reçu.
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu


_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu


_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
<Prev in Thread] Current Thread [Next in Thread>