ADSM-L

Re: [ADSM-L] hypothetical situation with dedup turned on

2012-11-16 10:55:20
Subject: Re: [ADSM-L] hypothetical situation with dedup turned on
From: "Prather, Wanda" <Wanda.Prather AT ICFI DOT COM>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Fri, 16 Nov 2012 15:51:11 +0000
I can confirm that if you do audit volume fix=yes , then a MOVE DATA, that it 
LOOKS like it works.  
The issue you mention below hadn't occurred to me. 
Ick.  Could turn ME into a vodka drinker..
But believe me, I"ve got "deduprequiresbackup yes" in place, and back up to a 
non-deduped copy pool.

(And, by the way, the term "full incremental" makes me twitch, even without the 
vodka!)



-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of 
Alex Paschal
Sent: Friday, November 16, 2012 1:09 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: [ADSM-L] hypothetical situation with dedup turned on

Hi, David.  You can still do as you're already doing:  "audit volume fix=yes" 
to find the damaged blocks, then do a "move data" against the good data.  That 
would leave the unreadable data on the volume.

If the copypool volume is unavailable for a "restore volume", then the only 
thing you could do is "delete volume discarddata=yes" and take the concomitant 
loss of data that refers to the bad blocks. TSM should then re-back up that 
data during the next full incremental backup.  (Full incremental?  Oxymoron!  
Also, maybe too much vodka. Stoli's Orange, tonight.  ;-)

Question for the IBMers:  Is TSM smart enough to delete all of the file objects 
that refer to the deduped/damaged/discarded blocks?  I would expect so, 
especially with the ~new DB2 referential integrity enforcement, but I think 
that's exactly what David's question is getting at.  Could we get an 
authoritative answer on that?

And a more egg-head question from me:  if a few damaged blocks are inside an 
aggregate, my understanding is that the entire aggregate would be marked bad 
during the audit, which means TSM wouldn't be able to move data 
reconstruct=yes, which would cause a larger footprint of data loss.  Is my 
hypothesis correct?

Hmm.  Now that I think about it, CRC would have to be enabled on the stgpool to 
detect those few bad blocks within an aggregate, otherwise the headers/magic 
numbers for the aggregate/blocks would still be readable/good and the aggregate 
would audit as intact. Thoughts?

Another question:  do file volumes get magic numbers?  Haha  (Sorry, I blame 
the vodka.)


On 11/15/2012 12:58 PM, Tyree, David wrote:
> This a hypothetical situation.
>
> In this situation the needed tape from the copy pool is not available.
> I realize that the data would be lost but how what you do next?
>
> if we were still running v5 of TSM we would do a move data (MOVE VOL XXX) to 
> save what we could then delete the volume (DEL VOL XXX). We would lose some 
> data but the next backup cycle would rebackup any missing active data.
>
> Since we are now running v6 with dedup it seems like the process would be 
> different. Each volume no longer contains a complete set of files. They now 
> contain parts of files.
>
>
> David Tyree
> Interface Analyst
> South Georgia Medical Center
> 229.333.1155
>
> -----Original Message-----
> From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf 
> Of Grigori Solonovitch
> Sent: Thursday, November 15, 2012 2:58 PM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Re: [ADSM-L] hypothetical situation with dedup turned on
>
> Have you tried to use standard copy pol to recover any problems in primary 
> pool?
>
> Grigori G. Solonovitch
> Senior Technical Architect  Ahli United Bank Kuwait  
> www.ahliunited.com.kw
>
> Please consider the environment before printing this E-mail
>
>
> -----Original Message-----
> From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf 
> Of Tyree, David
> Sent: Wednesday, November 14, 2012 10:36 PM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: [ADSM-L] hypothetical situation with dedup turned on
>
>                  I've had some sys admins ask me about a possible situation 
> with using dedup on our primary storage pool. We are currently using dedup 
> and I can't come up with a good answer.
>                  Ok, our primary storage pool is using dedup. Something 
> (corruption, whatever) happens to one of the files in the primary pool and 
> the data needed to recover the file in the primary pool is not available.
>                  I attempt to do a restore of the corrupt  file and the 
> needed tape is not available.
>                  How would I go about fixing that kind of a situation?
>                  Back before we started using dedup we could just do a move 
> volume to save what we could and then do a delete volume and the next backup 
> of the server would straighten everything out. We might lose inactive copies 
> but the next backup cycle would catch the missing active files.
>                  With the way dedup works I'm not sure what we would do.
>                  Any suggestions?
>
> David Tyree
> Interface Analyst
> South Georgia Medical Center
> 229.333.1155
>
>
> Please consider the environment before printing this Email.
>
> CONFIDENTIALITY AND WAIVER: The information contained in this electronic mail 
> message and any attachments hereto may be legally privileged and 
> confidential. The information is intended only for the recipient(s) named in 
> this message. If you are not the intended recipient you are notified that any 
> use, disclosure, copying or distribution is prohibited. If you have received 
> this in error please contact the sender and delete this message and any 
> attachments from your computer system. We do not guarantee that this message 
> or any attachment to it is secure or free from errors, computer viruses or 
> other conditions that may damage or interfere with data, hardware or software.
>