08-19-2014 08:30 AM
Hiii experts
NBU master 7.5.0.6 OS window 2003, NBU appliance 5220 2.5.6, I have a situation in my NBU setup. In GUI -- media -- various media's have difference between 2 columns images and valid images. Noticed this problem recently. I only perform tape copy/duplication on them for long term retention.
Single retention images saved on media. No data on media has expired. duplication job failed while doing tape copy on these 2 media.
taking example of 2 media's 1065L4 and 1062L4. Kilobytes , Images, Valid Images. differ from output of bpimmedia and nbemmcmd.
Images on media means: number of images written on media(successful+failed), Valid images means: number of successfull images written on media. Do netbackup clear or remove those failed images from media ?? Please share descriptive explaination
C:\Program Files\VERITAS\NetBackup\bin\admincmd>bpimmedia -U -tape -mediaid 1065L4
--------------------------------------------------------------------------------
Backup-ID: NAS1_1406955603
Policy: NDMP_Vol1_SnapM
Schedule Type: FULL
Retention Level 1
Number of Files: 10
Compression: N
Encryption: N
Image Type: Regular
Primary Copy: 2
Expires: 07:00 09/02/2014
Image on Hold: 0
Indexing Status: 0
Copy Number: 3
Fragment Number: IDX
Fragment Size (KB): 4
Media Type: RMed
Density: hcart
File Number: 2
Offset: 13412034
Host: mediaserver3
Device Written On: 2
MPX: -
Expires: -
Retention Level: -
MediaID: 1065L4
Copy on Hold: 0
Copy Number: 3
Fragment Number: 3
Fragment Size (KB): 858369920
Media Type: RMed
Density: hcart
File Number: 1
Offset: 2
Host: mediaserver3
Device Written On: 2
MPX: -
Expires: -
Retention Level: -
MediaID: 1065L4
Copy on Hold: 0
C:\Program Files\VERITAS\NetBackup\bin\admincmd>bpimmedia -U -tape -mediaid 1062L4
--------------------------------------------------------------------------------
Backup-ID: NAS2_1406948411
Policy: NDMP_VolSolaris_SnapM
Schedule Type: FULL
Retention Level 1
Number of Files: 7
Compression: N
Encryption: N
Image Type: Regular
Primary Copy: 2
Expires: 05:00 09/02/2014
Image on Hold: 0
Indexing Status: 0
Copy Number: 3
Fragment Number: IDX
Fragment Size (KB): 4
Media Type: RMed
Density: hcart
File Number: 2
Offset: 525387
Host: mediaserver1
Device Written On: 3
MPX: -
Expires: -
Retention Level: -
MediaID: 1062L4
Copy on Hold: 0
Copy Number: 3
Fragment Number: 2
Fragment Size (KB): 67249024
Media Type: RMed
Density: hcart
File Number: 1
Offset: 2
Host: mediaserver1
Device Written On: 3
MPX: -
Expires: -
Retention Level: -
MediaID: 1062L4
Copy on Hold: 0
NBEMMCMD, Version: 7.5.0.6
C:\Program Files\VERITAS\NetBackup\bin\admincmd>nbemmcmd -listmedia -mediaid 1065L4
====================================================================
Media GUID: a8664cfe-ea61-44b2-a98c-1b00ffd2462b
Media ID: 1065L4
Partner: -
Media Type: HCART
Volume Group: 000_00001_TLD
Application: Netbackup
Media Flags: 1
Description: ---
Barcode: 011065L4
Partner Barcode: --------
Last Write Host: mediaserver1
Created: 08/07/2014 14:13
Time Assigned: 08/13/2014 23:31
First Mount: 08/08/2014 12:50
Last Mount: 08/16/2014 16:24
Volume Expiration: -
Data Expiration: 08/22/2017 07:00
Last Written: 08/15/2014 00:38
Last Read: -
Robot Type: TLD
Robot Control Host: MASTER
Robot Number: 1
Slot: 10
Side/Face: -
Cleanings Remaining: -
Number of Mounts: 7
Maximum Mounts Allowed: 999
Media Status: FULL
Kilobytes: 1360165508
Images: 3
Valid Images: 2
Retention Period: 24
Number of Restores: 0
Optical Header Size Bytes: 1024
Optical Sector Size Bytes: 0
Optical Partition Size Bytes: 0
Last Header Offset: 13412037
Adamm Guid: 00000000-0000-0000-0000-000000000000
Rsm Guid: 00000000-0000-0000-0000-000000000000
Origin Host: NONE
Master Host: MASTER
Server Group: UNRESTRICTED_SHARING_GROUP
Upgrade Conflicts Flag:
Pool Number: 32
Volume Pool: Madrid
Previous Pool Name: scratch
Vault Flags: -
Vault Container: -
Vault Name: -
Vault Slot: -
Session ID: -
Date Vaulted: -
Return Date: -
Media on Hold: 0
====================================================================
Command completed successfully.
NBEMMCMD, Version: 7.5.0.6
C:\Program Files\VERITAS\NetBackup\bin\admincmd>nbemmcmd -listmedia -mediaid 1062L4
====================================================================
Media GUID: ca090c33-e025-4116-a25b-37118310e164
Media ID: 1062L4
Partner: -
Media Type: HCART
Volume Group: ---
Application: Netbackup
Media Flags: 1
Description: ---
Barcode: 011062L4
Partner Barcode: --------
Last Write Host: mediaserver3
Created: 08/07/2014 14:13
Time Assigned: 08/07/2014 01:48
First Mount: 08/08/2014 12:51
Last Mount: 08/11/2014 19:36
Volume Expiration: -
Data Expiration: 08/24/2017 12:46
Last Written: 08/11/2014 13:29
Last Read: -
Robot Type: NONE
Robot Control Host: -
Robot Number: -
Slot: -
Side/Face: -
Cleanings Remaining: -
Number of Mounts: 3
Maximum Mounts Allowed: 999
Media Status: FULL
Kilobytes: 1854468292
Images: 4
Valid Images: 2
Retention Period: 24
Number of Restores: 0
Optical Header Size Bytes: 1024
Optical Sector Size Bytes: 0
Optical Partition Size Bytes: 0
Last Header Offset: 0
Adamm Guid: 00000000-0000-0000-0000-000000000000
Rsm Guid: 00000000-0000-0000-0000-000000000000
Origin Host: NONE
Master Host: MASTER
Server Group: UNRESTRICTED_SHARING_GROUP
Upgrade Conflicts Flag:
Pool Number: 32
Volume Pool: Madrid
Previous Pool Name: scratch
Vault Flags: -
Vault Container: -
Vault Name: -
Vault Slot: -
Session ID: -
Date Vaulted: -
Return Date: -
Media on Hold: 0
====================================================================
Command completed successfully.
08-19-2014 09:09 AM
Number of images on media is the highest number of , non expired backups that were on the media
For example, if I write 4 separate images to a media, this will be 4
Number of Valid images = number of non-expired images,
So I write 4 successful backups,
No Images = 4
Valid Images = 4
As soon as one expires,
No images = 4
Valid images = 3
If an image to tape fails, NBU will rewind to the end of the previous image and write a new logical end of data mark, so although the failed image is not overwriteen at that time, it is effectivly inaccessable, as the tape drive cannot position past the logical end of data. The next time the media is written to, that failed image would be overwritten.
I'd have to run a test to check, but I think we only update no images on the tape when the write to the tape has completed (I think) so, in that case Images on media would only be the total number of successful images, not the total number of successful and failed, I'd need to check that though.
I see the media is full, I see also that each contains 1 minage (bpimmedia) but nbemmcmd shows 2.
So, I think the following has happened :
The media was written, it became full and the last image spanned across to another media, and then once writing to the next media it failed. We don't then reload the previous media into the drive to re-write the logical end-of-data mark, we just leave it but I suspect we do adjust the number of images in the media database, hence why you see a mis-match at least for media 1065L4, which shows 4 images and 3 valid, if the last one spanned and failed, only 3 images are valid, although the beginnin of the 4th image on that tape was successful (it only failed when it spanned).
This theory doesn't work for 1062L4 though, bpimmedia shows 1 images, nbemmcmd shows 4 images and 2 valid, if my theory is correct above, we can only be 1 image out, not 2 so it looks like there is a catalog inconsistency here.
So the only thing that can be done, is that you log a call and run NBCC. I would expect NBCC first run to produce a SRA file that requests mcontents to be run on these tapes so it can work out exactly what is on them, and show what the true state is. If my theory is wrong then 1065L4 would also be showing a catalog inconsistency, so NBCC would be required here as well.
08-19-2014 09:18 AM
OK, just checked - we only update the media DB with no images on tape once the backup completes, so for a media that does not become full (therefore no images span across to another media) then no of images = no of successful, non-expired images.
For a media that spans to another, we'll update media database when the tape becomes full, but if that backup / duplication fails on the next tape, I'm 99% sure we decrease the number of valid images (as that last image failed), but not the total number of images, as the part of the image that was written to the first tape was successful.
However, TBH this theory doesn't really matter as I still strongly recommend you run NBCC.
08-22-2014 07:30 AM
NBCC output uploaded for Case 0703363
08-25-2014 11:47 AM
The nbcc output is incomplete on the FTP server, I think some files are missing (depends which version you used) but some are 0 bytes, eg nbcc.gather
The NBCC should be zipped / gziped when it completes, just upload the one file as opposed to unzipping.
I have looked in a couple of the files for one of the tapes (so manually as opposed to running the NBCCA tool which I can't run without the correct files) ... and it does appear that there are inconsistances.
If you work with the TSE who has the case I'll keep an eye on it.
M
08-26-2014 03:32 AM
latest step done by me : NBCC output folder zipped then uploaded