CDL data to be moved to Data Domain Disk.

[Sid1987]

Hi Guys,

  Need a solution for a situation "I want to transfer or copy the data from CDL tapes to Data Domain Disk". As I have to Decom the CDL.

Please suggest me how can I achieve this, I have Netbackup 6.5.6 and 7.1.0.4. I believe bpduplicate will help me achieve it.

Please verify this. Also If I test this on a media id, How will I be able to verify if the Second copy has been created, I know its a lame question as a successful Duplicate job would suffice it , but still.

Also How can I validate the Data Domain image for restores.

And when I run the bpduplicate command or use GUI for entire media, will it be triggering a single job or number of jobs equivalent to the number of images in that media.

Thanks in advance

Sid

[Marianne]

What are your retention periods for VTL backups? 

Most people have short retention periods on VTL. If your retention periods are less than one month, why not just 'wait out' natural expiration of VTL and start doing backups to DD?

Use 'Media List' report to check retention/expiration dates.

[Sid1987]

Hi Marianne,

 Thanks for the reply. If they were to be expired then I would have let them. However many media's are retained for infinity(2038). So I have to move them. I tried for a backup id from a media. I saw a duplication job in activity monitor which completed. Now How do I verify things, like where can I or which command will give information about 2nd copy(duplicated copy). then I guess I will have to promote it as a primary copy to check the restore, how can I achieve that means which command and then after successful testing how can I revert it meaning change it back to secondary copy.

Thanks in advance.

Sid

[Marianne]

Apologies - I have never heard of VTL images kept for Infinity.

As I've said - all our customers using VTL keep images on VTL for 2 - 4 weeks (for quick restores) and duplicate to physical tape for long-term retention.

Easiest way to duplicate will be to duplicate by media-id instead of one-by one image, You can also select 'Primary' checkbox in the GUI to promote the new copy to primary.

The only REAL test is to perform restores from the duplicated copies. Depending on how many there are to test, you need to decide if successful duplication is good enough.... or just test restores from some of the duplicated images.

Have a look at these TN's and decide what will work best for you:

GUI : http://www.symantec.com/docs/HOWTO34391  (Method is same for Unix and Windows. Also see related links on this page)

cmd: http://www.symantec.com/docs/HOWTO43654  You can use '-set_primary' to automatically promote the new copy to primary.

You can start as many bpduplicates for media-id's that number of concurrent writes to DD will allow.

Source media/images can be expired if/when you are satisfied that duplications were successful.

[Sid1987]

Hi Marianne,

Thanks for the reply, I myself don't understand why they are there for infinity and we don't have physical tapes at all in our environment.

Could you tell me after testing duplicated images for restores how can I make them back secondary copy?

Thanks

Sid

[Sid1987]

Hi Marianne,

Could you tell me one thing, We have a situation. We have an SLP in which backups and replications are going to EDL(virtual tapes). Now we don't have any free space left in the Duplication EDL and we have a huge backlog for that SLP. If I register that media server to Data Domain with DDBOOST and use a DSU as Duplication destination in the SLP, Will the backloged duplication run on the new DSU?

Please reply.

Thanks in advance

Sid

[Mark_Solutions]

When an image is written using an SLP its destinations are "hard coded" into that image ... what i mean is that if the backup goes to device 1 and duplication to device 2 then even if you edit your SLP to now send the duplications to device 3 those images that were backed up to device 1 prior to you editing the SLP will still want to go to device 2.

So you need to either clear down space on your device 2 to allow the duplications to run or you will need to do a SLP cancel (nbstlutil command) to cancel the lifecycle of all outstanding images and then manually duplicate them to device 3.

Hope this makes sense and helps - do keep a careful eye on it all and save out lists of all affected images before you do anything so that you know exactly what has and what has not been affected.

The nbstlutil command will do most of it for you - check out the commanda guide for full details

In answer to your previous question - if you search the catalog against your Copy 1 original images and then sort by the Primary column so that the "no"'s show at the top, you can right click them and "make primary" again

Hope this helps

[Sid1987]

Hi Mark,

 I think there is a confusion here, I completely understand your explaination. However I Created a DSU and changed the Duplication to go to in the backlogged SLP. And guess what the duplication triggered and they started writting with Destination media id as @aaaid or something which I am sure is going to new data domain DSU.

Please suggest. Is it a bug or what?

Thanks

Sid

[Mark_Solutions]

OK - please clarify exactly what you have done ... am i assuming that you had an SLP with backup as VTL and Duplication as DD with a backlog of images to duplicate but the DD was full ....

You then created a DSU (where and on what was this created?) and edited the SLP, replacing the DD with this DSU and the backlog started to run to this new DSU?

Please clarify this and i will have a think about what has happened here

Thanks

[Sid1987]

Ok Mark,

 I had an SLP with backups going to a DD (VTL means the storage unit is a media manager type) and duplication going to EDL(VTL again with storage unit as a media manager type). Now as we didn't have much free space left on the EDL. I created a DSU on a different DD unit(other than the backups DD) which is a disk type storage unit through DDBOOST. And I edited the SLP to go to that DSU. Actualy I removed the EDL Storage unit from Storage Unit group and added the DSU and the SLP using that storage unit group now has only that DSU as duplication destination. I understand the new backups will go to the new DSU for replication, However when I edited the Storage unit group, 1st thing when Duplication triggered around 12 to 14 jobs came active and started writing on media id @aaaa7 which is for DSU's, now by any chance it is possible that I got 12 tp 14 completed backups by that time and they are those duplications running, and moreover if they are not running on new DSU they should be failing because the Storage unit group doesn't know about that EDL storage unit anymore.

Please suggest.

Thanks

Sid

[Mark_Solutions]

Take a look at the detailed information in the duplication tab - where it shows the media id as @aaaa7 - it should say a bit more to show where it is actually wrting to as well as the resources being requested.

See if it tells you a bit more, but yes @aaaa7 indicates a disk backup.

It may be that it is the use of the STU Group that got you around it as it is still duplicating to the group as per the original SLP but the group now has different things available

[Sid1987]

Hi Mark,

I forgot to mention a simple thing it is showing the new DSU besides the media @aaaa7. So we are sure they are going to DSU. Your explanation of that there is no change in storage unit group is reasonable, however could you confirm me that it is the valid reason for this thing happening.

Thanks

Sid

[Mark_Solutions]

If you have added a new member to the STU Group defined in the SLP then the SLP effectively has not changed and it can happily use the storage units within that group - sounds like what you have is working OK

If it was wrong then the old images would keep trying to go to the old umit, even though it did not exist.

To check that no images have been left behind just run:

nbstlutil stlilist -image_incomplete -l

This will list any images whose lefecycle is incomplete - if there are old ones in there then youy may need to dog deeper - if not then you are all good!

[Sid1987]

Hi Mark,

 It seems they are running, However I found many Duplicate jobs failed with error 191, though 191 is not the actual error its generic. Detailed status shows error code 14 and 12.

Here is the piece of it.

08/23/2012 03:55:47 - Info bpdm (pid=19595376) reading backup image
08/23/2012 03:55:47 - Info bptm (pid=19595376) using 256 data buffers
08/23/2012 03:55:47 - Error bptm (pid=19595376) unable to write to  /usr/openv/netbackup/db/config/shm/shi004pdmp02.shdc.chrysler.com_1345385677_copy1_2 file, No space left on device
08/23/2012 03:55:47 - Info bptm (pid=19595376) EXITING with status 14 <----------
08/23/2012 04:25:48 - Error bptm (pid=15007798) Could not open file /usr/openv/netbackup/db/config/shm/shi004pdmp02.shdc.chrysler.com_1345385677_copy1_2 to get shared memory information. Errno = 4: Interrupted system call
08/23/2012 04:25:50 - Error bpduplicate (pid=18704) host shi008nblp01.shdc.chrysler.com backup id shi004pdmp02.shdc.chrysler.com_1345385677 read failed, file write failed (14).
08/23/2012 04:25:50 - Info bptm (pid=15007798) EXITING with status 12 <----------
08/23/2012 04:25:50 - Error bpduplicate (pid=18704) host shi008nblp01.shdc.chrysler.com backupid shi004pdmp02.shdc.chrysler.com_1345385677 write failed, file open failed (12).
08/23/2012 04:25:51 - Error bpduplicate (pid=18704) Duplicate of backupid shi004pdmp02.shdc.chrysler.com_1345385677 failed, file open failed (12).
08/23/2012 04:25:51 - Error bpduplicate (pid=18704) Status = no images were successfully processed.
08/23/2012 04:25:53 - end Duplicate; elapsed time 17:40:20
no images were successfully processed  (191)

Could you please suggest me in what direction I should proceed, or I should straight away open a case with symantec.

Thanks

Sid

[Mark_Solutions · Accepted solution]

No space left on device? Is you NBU installation drive OK? Hasn't filled up has it?

[Sid1987]

Yes I checked actually the bptm logs were consuming /usr/openv. Now seems to be fine thanks.

Sid