We've got a library that holds some 326 media and its out of production,
so I just loaded them all up and let it run.  Phase I was somewhere in
the 3-4 day range.  Phase two was your normal 2-6 hours depending on
number of fragments etc...  We used to do a lot of multi-streaming so it
took a loooong time.  Thankfully it was only 5 or 6 restores.

-Jonathan 

-----Original Message-----
From: Justin Piszcz [mailto:jpiszcz AT lucidpixels DOT com] 
Sent: Monday, December 17, 2007 2:42 PM
To: Martin, Jonathan
Cc: Jon Bousselot; veritas-bu AT mailman.eng.auburn DOT edu
Subject: RE: [Veritas-bu] Media Tape lifespan

May I ask how long (time/duration) it took you to inventory all of those
tapes?

How long for phase I?
How long for phase II?

etc..?

Did you up-convert any of the data on the old media -> LTO-3?

If so, how did that go?

Justin.

On Mon, 17 Dec 2007, Martin, Jonathan wrote:

> Drives are Quantum SDLT 220s - Media is SDLT-I - I don't know the 
> media manufacturer off hand, I'd have to look.  We buy whatever is
cheapest.
> =P  Before being requested back the media was at our offsite / media 
> storage facility.  I'm not sure what kind of environmental stuff 
> they've got going on there.  To be honest, we don't keep data this 
> long per policy - but we did have several hundred media sitting around

> waiting to be destroyed and I was asked to inventory it.  So I guess a

> year or so in a cardboard box (in the plastic case) was part of the 
> life cycle as well.  Aside from what's being kept offsite for long 
> term retention we're 100% on LTO3 media now / all this media was going

> to be destroyed pending legal yadda yadda...
>
> -Jonathan
>
> -----Original Message-----
> From: Justin Piszcz [mailto:jpiszcz AT lucidpixels DOT com]
> Sent: Monday, December 17, 2007 2:07 PM
> To: Martin, Jonathan
> Cc: Jon Bousselot; veritas-bu AT mailman.eng.auburn DOT edu
> Subject: RE: [Veritas-bu] Media Tape lifespan
>
> Jonathan,
>
> That is quite remarkable and impressive!
>
> What type of tape media?
>
> How do you store the tapes?
>
> What type of environmental conditions?
>
> Had they not been used since 2001?/
>
> Justin.
>
> On Mon, 17 Dec 2007, Martin, Jonathan wrote:
>
>> I recently imported some 26,000 images from media going back to
> 7/2001.
>> I only did phase one, then produced a report.  We did phase two and 
>> restore on media going back to 1/2002 for windows file servers.
>>
>> -Jonathan
>>
>> Here's the technote you are refering to - I got some data back using 
>> this method.
>>
>> failed phase 2 import
>>
>> ACTION: Discovered that the image is on only one tape, so determine 
>> that the tape has a bad spot on it.  The phase 2 import ran for three

>> hours then gave the error "tar did not find all the files to be 
>> restored". to use the following procedure to read as much data off 
>> the
>
>> tape as possible.
>>
>> bpexpdate -m media_id -d 0
>> bpimage -cleanup -allclients
>>
>> Solution:
>> The following touch file allows for this to happen:
>> install_path\netbackup\db\images\"client_name"\xxxxxxxxxxxx\debug_add
>> _
>> fi
>> les
>> This touch file can be used under certain circumstances to restore 
>> from a partial import if the catalog is in a binary format. If the 
>> catalog is in an ASCII format, these steps can be used but you will 
>> not need to add the touch file.
>> The first phase of an import creates the header file (non.f). The 
>> second phase of an import creates the files file (.f file). If the 
>> import fails during the phase two import for some reason (bad tape 
>> block, missing tape in a multi-tape backup, or the backup was 
>> incomplete etc.), NetBackup will not be able to browse the image for 
>> restore even though both files needed are created (assuming that the 
>> second phase of the import did not fail right away).
>> 1. Perform the phase one import. Your directory structure will look 
>> something like this. The directory may have many images. This example

>> assumes there is only one image.
>> Directory of C:\Program
>> Files\VERITAS\NetBackup\db\images\Sap\1058000000
>> 07/22/2003 01:54 PM <DIR> .
>> 07/22/2003 01:54 PM <DIR> ..
>> 07/22/2003 01:51 PM <DIR> catstore
>> 07/22/2003 01:51 PM 822 Sap_test_1058889087_FULL 2. Create the 
>> debug_add_files touch file. Here is an example showing the new touch 
>> file.
>> Directory of C:\Program
>> Files\VERITAS\NetBackup\db\images\Sap\1058000000
>>
>> 07/22/2003 01:54 PM <DIR> .
>> 07/22/2003 01:54 PM <DIR> ..
>> 07/22/2003 01:51 PM <DIR> catstore
>> 07/22/2003 01:24 PM 0 debug_add_files
>> 07/22/2003 01:51 PM 822 Sap_test_1058889087_FULL 3. Run the phase two

>> import. As long as the phase two import gets to the point that 
>> actually starts writing entries in the files file, the structure may 
>> look like this.
>> Directory of C:\Program
>> Files\VERITAS\NetBackup\db\images\Sap\1058000000
>> 07/22/2003 01:54 PM <DIR> .
>> 07/22/2003 01:54 PM <DIR> ..
>> 07/22/2003 01:51 PM <DIR> catstore
>> 07/22/2003 01:24 PM 0 debug_add_files
>> 07/22/2003 01:51 PM 822 Sap_test_1058889087_FULL
>> 07/22/2003 01:51 PM 181,627 Sap_test_1058889087_FULL.f
>> 07/22/2003 01:50 PM 338,287 Sap_test_1058889087_FULL_debug.f 4. When 
>> the phase two import finally does fail or is stopped, rename the .f 
>> to
>
>> .f.old or any other name, and rename the _debug.f to just .f 
>> Sap_test_1058889087_FULL.f becomes Sap_test_1058889087_FULL.f.old 
>> Sap_test_1058889087_FULL_debug.f becomes Sap_test_1058889087_FULL.f
5.
>
>> This is where you would start if you are running an import to an 
>> ASCII
>
>> catalog and it fails.
>> Edit the header file (Sap_test_1058889087_FULL).
>> Change the "IMAGE_TYPE" from 1 to 2. The IMAGE_TYPE is set to 0 
>> (zero)
>
>> for backups, 1 after a phase one import and 2 after a successful 
>> phase
>
>> two import. If the IMAGE_TYPE is left at 1, Backup, Archive and 
>> Restore will not browse it for restores.
>> Example:
>> IMAGE_TYPE 2
>> Open the files file (Sap_test_1058889087_FULL.f) in a program that 
>> can
>
>> count line numbers. Textpad can do this by going to Configure | 
>> Preferences | View, and selecting the checkbox for Line Numbers. 
>> Count
>
>> the number of lines in the files file.
>> Edit the NUM_FILES entry in the header file to represent the number 
>> of
>
>> lines as above.
>> Example:
>> NUM_FILES 1778
>> You should then be able to open up Backup, Archive and Restore and 
>> browse the image for restore. Since this was a partial import, not 
>> all
>
>> files will be available for restore, but you may get the required
> file.
>>
>> -----Original Message-----
>> From: veritas-bu-bounces AT mailman.eng.auburn DOT edu
>> [mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of Jon 
>> Bousselot
>> Sent: Monday, December 17, 2007 1:32 PM
>> To: Justin Piszcz
>> Cc: veritas-bu AT mailman.eng.auburn DOT edu
>> Subject: Re: [Veritas-bu] Media Tape lifespan
>>
>> 6 years old, so probably written with NB 3.2 or 3.4.  AIT-2 media, 
>> five tapes, five fragments,  and fragment 4 was damaged in such a way

>> that NB could not finish the catalog import phase 2, but the data I 
>> wanted was on tape 3.
>>
>> Support gave me a tech doc that tells NB to keep the catalog data for

>> all images found prior to a failure.  Without this method, it would 
>> fail on the import and discard all the previous work. (several
>> hours!!)
>>
>> The AIT-2 media was used in a small tape changer that lived in an 
>> environment where drywall dust was present during construction, 
>> rendering a number of the tapes faulty.  Don't do drywall work in a 
>> live datacenter!!
>>
>>> Poll:
>>>
>>> What is the oldest tape you have restored from? 1 year? 3 years? 5+?
>>>
>>> Justin.
>> _______________________________________________
>> Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu 
>> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
>>
>

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu