[Pasig-discuss] Write-blocking a RAID

Cory Snavely csnavely at umich.edu
Fri Nov 2 13:03:28 EDT 2012 

This is the sort of request that I would fundamentally question. The 
primary concern with just doing what's asked is that a lot of the tools 
or methods one might use will not necessarily position the data well for 
long-term preservation, which is almost surely their objective.

I'm skeptical that they actually care about archiving the filesystem 
itself, but rather care about archiving its data contents, and are 
looking for a short cut for how to do that. It also might be the case 
that they don't want to do the diligence to e.g. extract file timestamps 
into more reliable/portable preservation formats like METS. That's 
somewhere between wrongheaded and dangerous, though. There really aren't 
short cuts in digital preservation...

All that said, assuming for the moment that there is some legitimate 
reason for doing this in this way, then don't forget about good old tar. 
 From the long-term digital preservation perspective, it is a stable, 
ubiquitously supported format, and there's every reason to think it will 
be around far into the future.

My $0.02.

On 11/01/2012 10:26 PM, Jacob Farmer wrote:
> Agreed, those are excellent suggestions. I particularly like #1 and #4,
> but the three key questions remain:
>
> 1)What operating system is the host?
>
> 2)What file system are the files sitting on?
>
> 3)Is the system broken
>
> a.If no, then you have lots of choices
>
> b.If yes, then what is the make and model of the RAID hardware and what
> is broken about it?
>
> *Jacob Farmer  |  Chief Technology Officer  |  Cambridge Computer  |
> "Artists In Data Storage" *
>
> Phone 781-250-3210  | jfarmer at CambridgeComputer.com
> <mailto:jfarmer at CambridgeComputer.com>  | www.CambridgeComputer.com
> <http://www.cambridgecomputer.com/>
>
> *From:*Chris Wood [mailto:lw85381 at yahoo.com <mailto:lw85381 at yahoo.com>]
> *Sent:* Thursday, November 01, 2012 8:37 PM
> *To:* Jacob Farmer
> *Cc:* jgpawletko at nyu.edu <mailto:jgpawletko at nyu.edu>;
> pasig-discuss at mail.asis.org <mailto:pasig-discuss at mail.asis.org>;
> digital-curation at googlegroups.com <mailto:digital-curation at googlegroups.com>
> *Subject:* Re: [Pasig-discuss] Write-blocking a RAID
>
> Good idea Jacob. Other ideas are:
>
> 1) mount file system RO.
>
> 2) export RAID volumes RO - assuming the array has this function
>
> 3) backup all data to a non-rewritable tape like Oracle Storage-tek
> Vol-safe tape media,
>
> 4) copy all data into a file system with WORM capability like Oracle SAM-QFS
>
> Pls feel free to email or call.
>
> Chris Wood
>
> (408) 218 7313
>
> Chris.wood at oracle.com <mailto:Chris.wood at oracle.com>
>
> Sent from my iPhone
>
>
> On Nov 1, 2012, at 12:55 PM, Jacob Farmer <jfarmer at cambridgecomputer.com
> <mailto:jfarmer at cambridgecomputer.com>> wrote:
>
>     Joe, what operating system and file system is the server running?
>
>     Also the email you forwarded (below) implies that the system might
>     be broken in some way. Is it up and running and you just want to
>     preserve it?  If so, I have software that can capture the file
>     system metadata and preserve that along with a hash of each file.
>     Then you could store the files themselves in any kind of storage
>     device.
>
>     -> Jacob
>
>     *Jacob Farmer  |  Chief Technology Officer  |  Cambridge Computer  |
>     "Artists In Data Storage" *
>
>     Phone 781-250-3210  | jfarmer at CambridgeComputer.com
>     <mailto:jfarmer at CambridgeComputer.com>  | www.CambridgeComputer.com
>     <http://www.cambridgecomputer.com/>
>
>     *From:*pasig-discuss-bounces at asis.org
>     <mailto:pasig-discuss-bounces at asis.org>
>     [mailto:pasig-discuss-bounces at asis.org
>     <mailto:pasig-discuss-bounces at asis.org>] *On Behalf Of *Joseph Pawletko
>     *Sent:* Thursday, November 01, 2012 2:55 PM
>     *To:* pasig-discuss at mail.asis.org
>     <mailto:pasig-discuss at mail.asis.org>;
>     digital-curation at googlegroups.com
>     <mailto:digital-curation at googlegroups.com>
>     *Subject:* [Pasig-discuss] Write-blocking a RAID
>
>     Hello, All.
>
>     I've been asked by some of our archivists to preserve the contents
>     of a RAID (and a server)
>
>     and I'm looking for write-blocking strategies so that the file
>     Modify-Access-Change data are not
>
>     modified.
>
>     I discussed this problem with Mark Matienzo, who gave me
>     permission to forward the email below
>
>     in case others are trying to solve the same problem.
>
>     Has anyone used a strategy different from those outlined by Mark?
>
>     Thank you for your time.
>
>     Best-
>
>     Joe
>
>     --
>
>     Joe Pawletko
>
>     Digital Library Technology Services
>
>     Bobst Library
>
>     New York University
>
>     ---------- Forwarded message ----------
>     From: *Mark A. Matienzo* <mark.matienzo at nyu.edu
>     <mailto:mark.matienzo at nyu.edu>>
>     Date: Tue, Oct 23, 2012 at 12:48 PM
>     Subject: RAID acquisition
>     To: Joe Pawletko <jgp at nyu.edu <mailto:jgp at nyu.edu>>
>
>     My understanding is that there seem to be two camps:
>
>     1) Acquire the contents of the array over a network, which requires
>     the system to be live. This option is obviously less ideal because
>     the contents of the system may change (logs, etc.), but doesn't
>     require you to know as much about what RAID level the array is running.
>
>     2) Acquire each drive individually, removing them from the array.
>     This obviously requires the system to be dead, but may add
>     considerably additional work since you need to reassemble the array.
>     This may be preferable if you know less about the system. There are
>     a few different applications that can help with this: RAID
>     Reconstructor <http://www.runtime.org/raid.htm>; X-Ways Forensics;
>     EnCase. The specific software will depend on whether it's a software
>     or hardware array.
>
>     I don't really know what's preferable in this case - I think it
>     largely depends on what you know about the server.
>
>     Mark
>
>     ----
>     To subscribe, unsubscribe, or modify your subscription, please visit
>     http://mail.asis.org/mailman/listinfo/pasig-discuss
>     _______
>     Save the Date - PASIG Dublin, October 17-19, 2012
>     _______________________________________________
>     Pasig-discuss mailing list
>     Pasig-discuss at mail.asis.org <mailto:Pasig-discuss at mail.asis.org>
>     http://mail.asis.org/mailman/listinfo/pasig-discuss
>
>
>
> ----
> To subscribe, unsubscribe, or modify your subscription, please visit
> http://mail.asis.org/mailman/listinfo/pasig-discuss
> _______
> Save the Date - PASIG Dublin, October 17-19, 2012
> _______________________________________________
> Pasig-discuss mailing list
> Pasig-discuss at mail.asis.org
> http://mail.asis.org/mailman/listinfo/pasig-discuss

More information about the Pasig-discuss mailing list