Veritas-bu

[Veritas-bu] LTO-2 LTO-3 mix [NC]

2007-04-26 08:27:11
Subject: [Veritas-bu] LTO-2 LTO-3 mix [NC]
From: misha.pavlov at sgcib.com (misha.pavlov AT sgcib DOT com)
Date: Thu, 26 Apr 2007 08:27:11 -0400
Bobby, but from what I understand:

> But if the bar code rule changes, how does NetBackup know that it is not 
a new tape?

By checking the media catalogs.

> If you have 25 tapes sitting in offsite storage that NB read as 
ABC001-025
> you pull them back for a restore, NetBackup will not know anything about 
them,
> assign them to scratch and you sit there wondering why the tapes don't 
load.

If you put them back in the robot for the restore, you have to make sure 
NBU knows about them.
If it does not - then you'd have to import them.
Also, NBU has rules on which tapes it can and which tapes it cannot 
overwrite ( ALLOW_MEDIA_OVERWRITE )


> Another scenario is that Quantum comes in and has to update the 
> firmware on a library and in the process has to remove all of the 
> tapes to perform the library calibration/verification (I have been 
> through several of these).  All the tapes are put back into the 
> robot (we know not all in the same place).
And it does not matter for NBU which slot has which tape - that's 
libraries business.

> How will NetBackup determine what is old, what is new, what is borrowed, 
which admin is blue.
Catalogs.

> I have worked on this for over 6 months and have had a case 
opened/closed with Veritas.
> They concur with my dooms day scenario.

They are there to solve the immediate issues mostly.
Scenarios are not welcome :-)


> I have already had a similar thing happen on the install of a new 
> library on the same master that read 8 characters instead of 6 on 
> the bar code label.  I caught it before tapes got overwritten.

I have no problems having 3 different media types in the 618 slots L700. 
All is peachy.


> Bobby Williams 
> 2205 Peterson Drive 
> Chattanooga, Tennessee  37421 
> 423-296-8200 
> 
> 
> From: veritas-bu-bounces at mailman.eng.auburn.edu [mailto:veritas-bu-
> bounces at mailman.eng.auburn.edu] On Behalf Of McCammont, Anderson (IT)
> Sent: Wednesday, April 25, 2007 11:30 PM
> To: bobbyrjw at comcast.net; Martin, Jonathan (Contractor); Spearman, 
> David; veritas-bu at mailman.eng.auburn.edu
> Subject: Re: [Veritas-bu] LTO-2 LTO-3 mix

> No - changing the media id generation rule (MEDIA_ID_BARCODE_CHARS 
> in vm.conf) affects the media id.  Changing the barcode rules 
> (potentially) changes the media type and volume pool.  Existing 
> tapes won't be reallocated to a new pool on a barcode change - if 
> they did then every inventory of the library would break netbackup. 
> vmupdate(1M) only assigns attributes to new media.
> 
> If we could put regexps into vmrule e.g.
> 
> (made up vmrule -listall -b)
> barcode          media    volume         max mounts/
> tag              type     pool             cleanings description
> 
------------------------------------------------------------------------------
> *CU$             HC2_CLN  None                    50 LTO Cleaning Tapes
> CLN              DLT2_CLN None                    20 DLT Cleaning 
> Tapes 
> *L4$             HCART4   Scratch                  0 LTO4 media
> *L3$             HCART3   Scratch                  0 LTO3 media
> *L2$             HCART2   Scratch                  0 LTO2 media
> <NONE>           DEFAULT  None                     0 
> <DEFAULT>        DEFAULT  Scratch                  0 ------  
> then the world would be a better place.  A trivial code change I'd 
> bet and a feature request that's been outstanding for several years.
> 
> 
> From: veritas-bu-bounces at mailman.eng.auburn.edu [mailto:veritas-bu-
> bounces at mailman.eng.auburn.edu] On Behalf Of Bobby Williams
> Sent: 25 April 2007 18:21
> To: McCammont, Anderson (IT); 'Martin, Jonathan (Contractor)'; 
> 'Spearman, David'; veritas-bu at mailman.eng.auburn.edu
> Subject: Re: [Veritas-bu] LTO-2 LTO-3 mix

> You can't change the bar code rules like that in mid stream.  It 
> would change the media id of every tape already in the library and 
> put every existing tape in scratch.
> 
> You are going to either have to partition the drives in separate 
> partitions (Yuk), or just pick your default tape type and add all of
> the other types manually (not a big deal).  Existing tapes don't use
> the bar code rules anyway, they are already known.  It is only NEW 
> to the robot tapes that need to have a bar code rule.
> 
> I have 2 i2K's with both LTO-3 and LTO-2.  No big deal.  I have them
> in the same partition.  I use the LTO-3's mostly for tape to tape 
> duplication for offsite (LTO-2 duplicated to LTO-3 works great). 
> Most of the clients can't push data fast enough to feed the LTO-3 
> tapes.  (Yes, I do multiplex some mid range x3 to feed tapes data). 
> I use Staging disks to catch the really slow ones and do disk to 
> tape staging to the LTO-3 tapes.  If you put the LTO-3's in their 
> own partition, you could not use any tape duplication between 
partitions.
> 
> Another major headache that I have at a site with 2 separate 
> libraries (both ADIC 1k's) is when one runs out of media, everything
> fails with 96 errors and then moves to the other physical library. 
> Not a big issue until failure stats are looked at.
> 
> 

> Bobby Williams 
> 2205 Peterson Drive 
> Chattanooga, Tennessee  37421 
> 423-296-8200 
> 
> 
> From: veritas-bu-bounces at mailman.eng.auburn.edu [mailto:veritas-bu-
> bounces at mailman.eng.auburn.edu] On Behalf Of McCammont, Anderson (IT)
> Sent: Tuesday, April 24, 2007 10:24 PM
> To: Martin, Jonathan (Contractor); Spearman, David; veritas-
> bu at mailman.eng.auburn.edu
> Subject: Re: [Veritas-bu] LTO-2 LTO-3 mix

> If you do this you need to take part of your six digit barcode range
> and dedicate it to the particular media type. 
> 
> It's sad that Netbackup doesn't support proper regexps in the 
> barcode rules so that the 'L2', 'L3' (presumably) 'L4' last two 
> digits of the barcode can be used to select the media type.  We've 
> had a feature request in for years for this and it's gone nowhere 
> (unless 6.x does it and our rep hasn't mentioned it).
> 
> <rant>  why's there no feedback on feature requests - even when they
> do make it into a release you're left to scrabble round to see if it
> made it in there.  </rant>
> 
> From: veritas-bu-bounces at mailman.eng.auburn.edu [mailto:veritas-bu-
> bounces at mailman.eng.auburn.edu] On Behalf Of Martin, Jonathan 
(Contractor)
> Sent: 25 April 2007 03:42
> To: Spearman, David; veritas-bu at mailman.eng.auburn.edu
> Subject: Re: [Veritas-bu] LTO-2 LTO-3 mix

> Use barcode rules and media types.  LTO2 media can be hcart2 and 
> LTO3 can be hcart3 (I think hcart stands for Half-Inch Cartridge.) 
> NBU will never put a hcart2 tape into an hcart3 tape device.  TO NBU
> that's as incompatible as having LTO and DLT in the same library.
> 
> -Jonathan
> 
> From: veritas-bu-bounces at mailman.eng.auburn.edu [mailto:veritas-bu-
> bounces at mailman.eng.auburn.edu] On Behalf Of Spearman, David
> Sent: Tuesday, April 24, 2007 3:17 PM
> To: veritas-bu at mailman.eng.auburn.edu
> Subject: [Veritas-bu] LTO-2 LTO-3 mix

> 
> We will be getting an expansion cabinet to our scaler 2000 soon. At 
> the moment we have LTO-2 drives and tapes. When we add the expansion
> cabinet we are also going to add 2 LTO-3 drives with a set of LTO-3 
> tapes. My question is how do I keep LTO-2 tapes going to LTO-2 
> drives while the LTO-3 tapes go only to LTO-3 drives. I suspect a 
> combination of storage units and tape pools but now that we are at 
> NBU6mp4 expired tapes just return to the scratch pool which would 
> make them available to anything.
> 
> present setup
> 
> W2K3 master/media with sso/ndmp to scaler 2000 (5 tape drives allowed)
> W2K3 media server sso to scaler 2000 and DSSU (4 tape drives allowed)
> W2K3 san media server sso to scaler 2000 
> 
> David Spearman
> County of Henrico
> 
> 
> 
> NOTICE: If received in error, please destroy and notify sender. 
> Sender does not intend to waive confidentiality or privilege. Use of
> this email is prohibited when received in error.
> 
> NOTICE: If received in error, please destroy and notify sender. 
> Sender does not intend to waive confidentiality or privilege. Use of
> this email is prohibited when received in error.
> _______________________________________________
> Veritas-bu maillist  -  Veritas-bu at mailman.eng.auburn.edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

*************************************************************************
This message and any attachments (the "message") are confidential and
intended solely for the addressees.
Any unauthorised use or dissemination is prohibited. 
E-mails are susceptible to alteration.   
Neither SOCIETE GENERALE nor any of its subsidiaries or affiliates 
shall be liable for the message if altered, changed or falsified. 

*************************************************************************

<Prev in Thread] Current Thread [Next in Thread>