Trouble backing up Shadow copy components for a DFS share

[contra04]

netbackup 7.0 on windows server 2003

client is server 2008 sp2

Hi we are trying to backup a dfs share.

The client is enabled in the master server properties

the File list is:

______________

D:\

Shadow Copy Components:\

______________

How to people manually specify certain dfs shares? it looks something like this but how can I enumerate manually?

it is something like \User?Data\Distributed File System Replication\DfsrReplicatedFolders\Users\

How do I learn HOW to write the path? where is it stored?

just curious, either way the backup fails, and has failed since 2009, as the previous backup dudes didnt know that dfs had to be treated differently. They were getting just the dfsrprivate directory and not the shares since the turned on dfs in feb 2009.  New york is not happy they have asked for a restore and we have no data......

When using VSS and not VSP do I need to exclude anything for antivirus?

any ideas?

16/09/2011 09:00:00 - requesting resource Quantum-PX502
16/09/2011 09:00:00 - requesting resource harry.NBU_CLIENT.MAXJOBS.grub-bak
16/09/2011 09:00:00 - requesting resource harry.NBU_POLICY.MAXJOBS.grub-bak_dfs
16/09/2011 09:00:00 - granted resource harry.NBU_CLIENT.MAXJOBS.grub-bak
16/09/2011 09:00:00 - granted resource harry.NBU_POLICY.MAXJOBS.grub-bak_dfs
16/09/2011 09:00:00 - granted resource BTR078
16/09/2011 09:00:00 - granted resource HP.Ultrium3-SCSI.000
16/09/2011 09:00:00 - granted resource Quantum-PX502
16/09/2011 09:00:00 - estimated 21036774 Kbytes needed
16/09/2011 09:00:00 - begin Parent Job
16/09/2011 09:00:00 - begin Snapshot, Start Notify Script
16/09/2011 09:00:00 - started process RUNCMD (8356)
16/09/2011 09:00:00 - started
16/09/2011 09:00:04 - ended process 0 (8356)
Status 0
16/09/2011 09:00:04 - end Snapshot, Start Notify Script; elapsed time: 00:00:04
16/09/2011 09:00:04 - begin Snapshot, Step By Condition
Status 0
16/09/2011 09:00:04 - end Snapshot, Step By Condition; elapsed time: 00:00:00
16/09/2011 09:00:04 - begin Snapshot, Stream Discovery
Status 0
16/09/2011 09:00:04 - end Snapshot, Stream Discovery; elapsed time: 00:00:00
16/09/2011 09:00:04 - begin Snapshot, Read File List
Status 0
16/09/2011 09:00:04 - end Snapshot, Read File List; elapsed time: 00:00:00
16/09/2011 09:00:04 - begin Snapshot, Create Snapshot
16/09/2011 09:00:04 - started process bpbrm (8832)
16/09/2011 09:00:13 - begin Create Snapshot
16/09/2011 09:00:44 - end Create Snapshot; elapsed time: 00:00:31
16/09/2011 09:00:48 - end writing
Status 0
16/09/2011 09:00:48 - end Snapshot, Create Snapshot; elapsed time: 00:00:44
16/09/2011 09:00:48 - begin Snapshot, Policy Execution Manager Preprocessed
Status 13
16/09/2011 12:16:42 - end Snapshot, Policy Execution Manager Preprocessed; elapsed time: 03:15:54
16/09/2011 12:16:42 - begin Snapshot, Stop On Error
Status 0
16/09/2011 12:16:42 - end Snapshot, Stop On Error; elapsed time: 00:00:00
16/09/2011 12:16:42 - begin Snapshot, Delete Snapshot On Exit
16/09/2011 12:16:42 - begin Snapshot, Delete Snapshot On Exit
Status 0
16/09/2011 12:16:42 - end Snapshot, Delete Snapshot On Exit; elapsed time: 00:00:00
16/09/2011 12:16:42 - begin Snapshot, End Notify Script
16/09/2011 12:16:43 - started process RUNCMD (9228)
16/09/2011 12:16:46 - ended process 0 (9228)
Status 0
16/09/2011 12:16:46 - end Snapshot, End Notify Script; elapsed time: 00:00:04
Status 13
16/09/2011 12:16:46 - end Snapshot, Delete Snapshot On Exit; elapsed time: 00:00:04
file read failed(13)

and when it retried:

16/09/2011 12:15:06 - requesting resource Quantum-PX502
16/09/2011 12:15:06 - requesting resource harry.NBU_CLIENT.MAXJOBS.grub-bak
16/09/2011 12:15:06 - requesting resource harry.NBU_POLICY.MAXJOBS.grub-bak_dfs
16/09/2011 12:15:10 - granted resource harry.NBU_CLIENT.MAXJOBS.grub-bak
16/09/2011 12:15:10 - granted resource harry.NBU_POLICY.MAXJOBS.grub-bak_dfs
16/09/2011 12:15:10 - granted resource BTR078
16/09/2011 12:15:10 - granted resource HP.Ultrium3-SCSI.000
16/09/2011 12:15:10 - granted resource Quantum-PX502
16/09/2011 12:15:11 - estimated 0 Kbytes needed
16/09/2011 12:15:41 - connecting
16/09/2011 12:15:45 - connected; connect time: 00:00:04
16/09/2011 12:15:45 - begin writing
16/09/2011 12:16:39 - Error bpbrm(pid=4992) socket read failed, An existing connection was forcibly closed by the remote host.  (10054)
16/09/2011 12:16:39 - Error bptm(pid=9516) socket operation failed - 10054 (at child.c.1294)      
16/09/2011 12:16:39 - Error bpbrm(pid=4992) could not send server status message       
16/09/2011 12:16:39 - Error bptm(pid=9516) unable to perform read from client socket, connection may have been broken
16/09/2011 12:16:42 - end writing; write time: 00:00:57
file read failed(13)

[J_H_Is_gone]

been there done that gave back the t-shirt.

First - the dfs files do not show up in the drive backup when in the BAR but under the SCC

So if you went to the bar and then to drive E and looked for your usrs dir you would not see it

but go to

Shadow Copy Components:\User Data\Distributed File System Replication\DfsrReplicatedFolders\usrs

and look for your file here

---------------

2 options -  (first note you should have DFS and previous versions on - causes issues with dfs)

Option 1 -  you can turn dfs off and do a normal backup of the drives - this will get all the data.  this is normally done with the bp start notify and bp end notify scripts.

Option 2 - has 2 ways

Option A - the way it should be working as stated above the files get backed up in the SCC job under User Data

Option B - you can specify the dirs by using that path I show above

Issue with B is if you have more than one DFS dir with the same name:

Example:  E:\NY\usrs and E:\TX\usrs

if you specify the long path I gave you you will only get NY and not TX

you have to go into the registry to get the unique names.

Reading references:

TECH71076

TECH148740

Also you have to make sure you have enough room for the Shadow copy.

at command line on the windows dfs server

vssadmin List ShadowStorage

it should be at LEAST 10% of the size of the DFS files on the server.

and it will use VSS

[contra04]

Thank you very much for your reply: If option one just turns everything back to normal, Ill try it tonight.

Option 1 -  you can turn dfs off and do a normal backup of the drives - this will get all the data.  this is normally done with the bp start notify and bp end notify scripts.

Ok going to go with this option - I have added it to bpstart/end notify

do I need this hotfix? I am using netbackup 7.0 - or is this hotfix only for people who are going to do a shadow copy backup while the dfsr service is still running.

http://www.symantec.com/business/support/index?page=content&id=TECH64620

I am not sure what you mean by this:

First - the dfs files do not show up in the drive backup when in the BAR but under the SCC

what is the BAR? If I go with option 1, will the files just appear as normal folders under the D:\ - how do I navigate to them ?

I am going to do this on backup exec aswell, because it fails there too, on the server in NY.(crashes the VSS writer) Netbackup is used on the UK server.

[J_H_Is_gone]

BAR (Backup Archive and Restore window)

When you use the BAR and choose source and destination server

then say Select for restore

then choose your dates to restore

When you browse in the lower left screen - if DFS was running when the backup was done the files will not appear if you follow your tree down the E drive.

you have to follow the tree down the SCC backup to find the DFS files that were backed up.

but if you do option 1 and DFS is OFF when you do the backup - then yes the files will appear under the Drive Letter when you navigate the BAR.

your link does not point me to a specific Tech doc - so cannot answer that.

and you did add the commands to stop and start DFS to the bp notify scripts - right?

[contra04]

yes to bpstart_notify, bpend_notify, on the server to be backed up.

Thank you for your help with this.

It worked !

same on backup exec worked in NY also.

Both sides of the pond looking good.  Still a shame that I couldnt get the VSS writers to work on either.

[J_H_Is_gone]

if the writers are not in a stable state - the only way I know to fix them is a reboot.

You also have the issue that if you do the backup via SCC using VSS it backs it up as one big job.  I have a server that is so large that it takes 4 days to backup - so I have to find some other way to back it up as we do NOT want to stop/start DFS for the backups.

Be sure to check with the windows admins to make sure that DFS is not having any issues with you stopping and restarting it - and Make sure that Previous Versions is NOT running on the DFS server.

[contra04]

I am the windows admins :) - and it is causing a problem we have a DFS backlog now, as we only alloq 128k between 8pm and 1am for replication.

  Would really like the VSS to work. However the backupjob was hanging the writers.

[J_H_Is_gone]

 Depending on how big the DFS dirs are it can take a while for the VSS to build the shadow copy.

I have one that can take a day or more on the full backup just to build the shadowcopy before it ever starts backing up to tape.

You only have two choices - turn it of and do a flat file backup. which can cause your DFS to get behind when you restart it

or do the backups via the Shadow Copy Compotes job - you can break that up into more than one job - but as I am still fighting this issue I would suggest you call support and get some help in trying to get it backed up in a timely manner.