cancel
Showing results for 
Search instead for 
Did you mean: 

Backup skipping sub folder

Wes_T__Gator
Level 5

I am experiencing a similar issue as this post: https://www-secure.symantec.com/connect/forums/netbackup-76-backup-data-issue-error-90 I have a policy that backs up 4 clients at 4 different locations.  It's backing up the same path for each site.  Two of the four backups are working properly, the other two are backing up the folder I am specifying, but nothing below it.  We've spent hours checking the policy, client configs, and even the accounts the NBU services are using on the clients, but haven't come up with anything.  I was going to open a support ticket on Monday, but if anyone has a suggestion of something I can try now, I'm all ears.  I've attached the most recent bpbkar from one of the clients that is not backing up as expected.

1 ACCEPTED SOLUTION

Accepted Solutions

Wes_T__Gator
Level 5

After having the Windows Admin look at it again to figure out why one server was working and another server was not, he did find an entry in the DFS management for the two servers that were having issues.  DFS was disabled, so he assumed that it wasn't affecting anything, but I asked him to go ahead and delete them anyway.  

We made several changes, including restarting DFS services on at least one of the servers.  I also added a Full schedule with Accelerator forced rescan enabled.  I have kicked off another backup on both of the clients that were giving us trouble before.  They seem to be working ok, but I won't know for certain until the backup finishes and I see what is available for restore.  Per your suggestion Marianne, I dropped the logging down to 3. ;)

View solution in original post

14 REPLIES 14

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified
Have you checked if this folder is under DFSR control? Any chance that you can upload bpbkar as .txt extension? I cannot open zip files on a mobile device.

Nicolai
Moderator
Moderator
Partner    VIP   

I look like you are running 7.6, you should patch to 7.6.0.4

Could it be there is junction points involved ?

How to backup data within a junction point or symbolic link?

http://www.symantec.com/docs/TECH72326

I can't see any status 90 i the log file - did you upload the right one ?

I do see this message - But I do not know if its something to care about or not. Severity is <4> so it should be OK to ignore.

JBD - not using change journal data for <E:\Users\>: unable to validate change journal usage <reason=previous backups do not contain valid USNs>

 

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Pity that you are looking for assistance over a weekend and then do not respond to attempts to help....

I had a look at the bpbkar log. 

There is no failure logged in the file.
Just a successful incremental backup of 5120 bytes for E:\Users.

Please post a bpbkar log where status 90 was seen.
Ensure logging level is 1 or higher (3 should be fine.... Symantec support will ask for level 5).
Level 0 logs does not show file-level entries.

Wes_T__Gator
Level 5

My backups were not failing with a status 90.  My issue was similar, but not exactly the same.  I spoke to my Windows admins and they told me that the folder I was backing up was not using DFS, but for the two servers which were failing, I changed the Backup Selections to ALL_LOCAL_DRIVES and they were successful.   There were a couple of different posts/websites that got me to that point, but thanks to Marianne for putting me on the right path!  

Sorry for not being able to update the status over the weekend, but until the backups finished, I didn't have anything to add.  Now I just have to go back and exclude all the extra stuff that I don't need that was included in the A_L_D backup.

Does anyone know if using Shadow Copy Components:\ instead of A_L_D would work?   I want to make sure that I get the data that I need, but I would rather not back up more than I need to, especially since these are remote sites with very slow connectivity.  

 

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified
We need slightly higher logging level to see what is going on. We cannot say without seeing how volumes and folders are getting evaluated at backup time. Before next full backup, change policy to ALL_LOCAL_DRIVES again and select Allow multiple data streams. This will create a new job for each volume. In the meantime, have a look at bpmount output on this client.

Wes_T__Gator
Level 5

I up'd the logging to 5 before kicking off the backup Friday evening.  I've attached the logs below for one of the clients.

I changed it to A_L_D before kicking off the last backup on Friday, so the logs will reflect that.  One of the clients lost power over the weekend and restarted but the other seems to have completed successfully.  It looks like the folders I was trying to backup via the E drive were backed up via the Shadow Copy components.   I've kicked off an Incremental backup for the client who's full backup completed successfully.

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified
Sorry. Hopefully someone else on Connect has time for level 5 logs. I can look at up to a level 3 log. Curious to see output of bpmount.

Wes_T__Gator
Level 5

I know the logs are big, but I was anticipating having to open a ticket with support this week, so I wanted to have the logs they were going to ask for already.

C:\Program Files\Veritas\NetBackup\bin>bpmount
local: Shadow Copy Components:\ on Shadow Copy Components:\
CDROM: A:\ on A:\
local: C:\ on C:\
CDROM: D:\ on D:\
local: E:\ on E:\
EXIT STATUS 0: the requested operation was successfully completed

Here's the output from bpmount.

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Nothing wrong with bpmount output.

I honestly hope someone else on Connect has the time to look at your bpbkar log.
(I will be on leave over the next 2 weeks and will just pop in every now and then on my tablet...)

Wes_T__Gator
Level 5

It looks like Accelerator doesn't work with the Shadow Copy backup selection.  Is that correct?  I'm seeing this in the backup details:

03/18/2015 02:00:16 - Info bpbrm (pid=22394) reading file list for client
03/18/2015 02:00:19 - Info bpbrm (pid=22394) accelerator enabled
03/18/2015 02:00:25 - connecting
03/18/2015 02:00:28 - Info bpbrm (pid=22394) starting bpbkar on client
03/18/2015 02:00:28 - connected; connect time: 0:00:00
03/18/2015 02:00:31 - Info bpbkar (pid=1832) Backup started
03/18/2015 02:00:31 - Info bpbrm (pid=22394) bptm pid: 22450
03/18/2015 02:00:31 - Info bpbkar (pid=1832) change time comparison:<enabled>
03/18/2015 02:00:31 - Info bpbkar (pid=1832) accelerator enabled backup, archive bit processing:<disabled>
03/18/2015 02:00:31 - Info bptm (pid=22450) start
03/18/2015 02:00:33 - Info bptm (pid=22450) using 262144 data buffer size
03/18/2015 02:00:33 - Info bptm (pid=22450) using 250 data buffers
03/18/2015 02:00:36 - Info bptm (pid=22450) start backup
03/18/2015 02:00:38 - Info bpbkar (pid=1832) not using change journal data for <Shadow Copy Components:\>: not supported for non-local volumes / file systems
03/18/2015 02:00:38 - Info bpbkar (pid=1832) not using change journal data for <System State:\>: not supported for non-local volumes / file systems
03/18/2015 02:00:38 - Info bpbkar (pid=1832) not using change journal data for <_BACKUP_SPECIAL_OBJECTS AFTER System State:>: not supported for non-local volumes / file systems
03/18/2015 02:00:59 - Info bptm (pid=22450) backup child process is pid 22593
03/18/2015 02:00:59 - begin writing

sdo
Moderator
Moderator
Partner    VIP    Certified

The term "Shadow Copy backup" is a bit misleading I think.  What you see in the log is a "Shadow_Copy_Components:" backup - which is always full, and not (AFAIK) "accelerated".  Same is true for the inner component of "Shadow_Copy_Components:" which is known as a "System_State:" backup.

As for a "shadow copy backup", well this could be any backup which is VSS based - and most other areas of Windows file system backups will very likely be "shadow copy" based, i.e. abstracted by VSS.

FYI - ALL_LOCAL_DRIVES includes Shadow_Copy_Components: which includes System_State:.  Therefore if specify ALL_LOCAL_DRIVES, the NetBackup Client for Windows will attempt a stream for Shadow_Copy_Components:, therefore you do not need to specify it again in a policy which already includes/selects ALL_LOCAL_DRIVES.

HTH.

sdo
Moderator
Moderator
Partner    VIP    Certified

FYI - further information:

https://msdn.microsoft.com/en-us/library/windows/desktop/bb891959(v=vs.85).aspx

I do know that NetBackup Client for Windows honours the "FilesNotToBackup" key, i.e. NetBackup queries the Windows registry so that it knows, for any given server and/or OS version, which files not to backup:

http://www.symantec.com/business/support/index?page=content&id=TECH128661

What I would like confirmed is whether NetBackup Client for Windows will still record an entry for files (and their sizes) listed in "FilesNotToBackup" BUT that in the event of a restore an empty but size pre-allocated file is created?  So, for example, whilst a page file may appear to have been backed-up, because it's name appears in the BAR GUI, it hasn't actually been backed-up, and has not consumed any 'backup storage' and no data was transfered.  Is this true?  If I get time, I'll try to test it.

Wes_T__Gator
Level 5

Sorry for the confusion, I was in a hurry and didn't type out the full Backup Selection, which was Shadow Copy Components:\.  

Wes_T__Gator
Level 5

After having the Windows Admin look at it again to figure out why one server was working and another server was not, he did find an entry in the DFS management for the two servers that were having issues.  DFS was disabled, so he assumed that it wasn't affecting anything, but I asked him to go ahead and delete them anyway.  

We made several changes, including restarting DFS services on at least one of the servers.  I also added a Full schedule with Accelerator forced rescan enabled.  I have kicked off another backup on both of the clients that were giving us trouble before.  They seem to be working ok, but I won't know for certain until the backup finishes and I see what is available for restore.  Per your suggestion Marianne, I dropped the logging down to 3. ;)