Yup, I"ve seen the same type of bottleneck, esp. on Windows file servers

Alternatives:

Check into creating image backups:  Restoring an image will take a lot
less time than restoring 10 million files individually (because you
don't go through the file create process for each individual file, as
Kelly said. ). Then you "roll forward" by restoring files from TSM
incrementals that are newer than the image.
Question is, can you find an optimum point where you can take the image
dumps often enough tp still get your complete restore time within your
SLA's.  Works best with data that has a very low change rate.

Hardware solutions: physical mirrors or hardware-type flashcopy for the
filesystem; then the same "roll forward" idea by restoring files from
TSM that are newer than the image.  Physical mirrors or flashcopy is
probably going to be the most satisfactory solution in the long run.

Software solutions: you can experiment with virtual mount points (if
this is a Unix system) in TSM; lets you back up the 1 filesystem as
several.  Then you can collocate by filesystem in TSM, use multiple
restore streams.  That may not help much, if the disk itself is the
bottleneck.

You have to find the solution that works for your particular OS, and
hardware.

The BEST solution, is to educate folks not to stick 10mil files in 1
filesystem, in the first place...

Wanda Prather
"I/O, I/O, It's all about I/O"  -(me)

  

-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Wallace.Dwight
Sent: Wednesday, May 17, 2006 8:33 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: Disk-to-Disk Backup

Another interesting thought.
We had thought about this one.




-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Kelly Lipp
Sent: Wednesday, May 17, 2006 3:12 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: Disk-to-Disk Backup

Before spending a ton of time optimizing this realize that the
impediment to fast restore is file create time on the file server.  My
testing has shown that we can create between 50K and 75K files per hour.
10 Million files at 50K files is a long time: 200 hours.  Won't matter
how many TB of disk pool you have... 


Kelly J. Lipp
VP Manufacturing & CTO
STORServer, Inc.
485-B Elkton Drive
Colorado Springs, CO 80907
719-266-8777
lipp AT storserver DOT com

-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Christoph Pilgram
Sent: Tuesday, May 16, 2006 7:43 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: [ADSM-L] Disk-to-Disk Backup

Hi all,

Because we have problems to hold our service level agreements with the
customers for restoring big file-servers (10 Mio files, 1TB disk-space
in one filesystem), we are thinking about storing the backups not
anymore on tape but on disk. Does anybody has experience with that kind
of storage-pool for about 40 TB of backup data ? Does anybody use for
example a "Data Domain DD460" or other systems using COS to reduce the
amount of data.

Thanks for help

Chris