Eric,
The fragmentation occurs because of aggregates.  The files within the
aggregates expire but the space isn't freed up until all the files in an
aggregate expire.
Jim Sporer


At 11:14 AM 6/2/2004 +0200, you wrote:
Hi Jim!
> I know that this will involve running reclamation on the ATA storage pool
but we felt using a DISK device class would waste to much storage because of
fragmentation.

Why did you think fragmentation was a waste of space? If a file expires, the
space it allocated on a diskpool becomes free space, so it shouldn't be
considered wasted.
I can imagine that when a lot of little files expire, one gets a lot of
fragmented free space, which could have impact on the specific disks
performance in the long run.
Funny to see that a lot of people (including us) are thinking about SATA
storage for their backups. There doesn't seem to be anybody with real
hands-on experience however...
Kindest regards,
Eric van Loon
KLM Royal Dutch Airlines


-----Original Message----- From: Jim Sporer [mailto:[EMAIL PROTECTED] Sent: Tuesday, June 01, 2004 21:37 To: [EMAIL PROTECTED] Subject: Re: D2D backup with TSM


We are going to try disk only backups using ATA storage. We've decided to use use a FILE device class for the ATA storage pool and do the backups directly to the ATA storage pool. We will back up the data to a copy storage pool defined on a 3494 tape library. I know that this will involve running reclamation on the ATA storage pool but we felt using a DISK device class would waste to much storage because of fragmentation. Jim Sporer University of Wisconsin

  At 10:05 PM 5/28/2004 -0400, you wrote:
>We recently had a presentation from EMC.  During this presentation
>they discussed their new DL700
>virtual tape library.  It got people talking about a long
>term strategy of moving to a all disk based backup system.  So we
>started thinking of the pros and cons of how to go about setting up TSM
>for a disk based backup system.
>
>Here are some initial thought on ways to configure TSM for D2D
>backups.  We would be very interested in your thoughts/comments.  I
>doubt we are the only people thinking about this topic.
>
>1)Purchase a very large disk system (ata drives?) and put storage pools on
>them.
>     - use a standard DISK device pool for backups
>         - how to reclaim space?  do you even need to?
>         - fragmentation problems?
>         - multiple node access concurrently
>     - use a FILE device based pool
>         - single node access per disk file volume
>         - need to run reclamation
>         - still stage to disk and migrate to FILE device based pool?
>     - use a tape copy pool for offsite and backup
>     - use a offsite disk pool somehow
>         - iscsi over lan or fc if close enough (enough throughput?)
>         - other?
>     - must use client compression to get data compressed
>         - we mostly don't do this today
>
>2)  Purchase a virtual tape system like the DL700
>     (bundle of a EMC Clariion and a FalconStor vts appliance)
>     - provides compression for data
>     - appliance is responsible for layout and use of disk space
>     - can copy a virtual tape to a real physical tape for offsite storage
>
>
>Thanks
>
>Rick


********************************************************************** For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachments, nor responsible for any delay in receipt. **********************************************************************

Reply via email to