cancel
Showing results for 
Search instead for 
Did you mean: 

Problem backing up Windows 2008 with DSFR on NBU 7.0

Gerald_W__Gitau
Level 6
Certified
My setup is as follows:

Master server - windows 2003 SP2 STD edition NBU version 7.0
Media server - windows 2003 SP2 STD edition NBU version 7.0
Client - windows 2008 R2 with DFSR replication NBU version 7.0 with DFSR fix applied (ET1940549)
Backup selection - ALL_LOCAL_DRIVES

When the backup kicks off, it does everything right and stays at begin writing......after 30 or so minutes the backup fails with either error 13 (mostly) or 24. Not a single byte is wriiten. The problem seems to be the shadow copy components backup that fail. I have tried selecting drives e.g C:\ and the backup works perfectly. I tried the directive shadow copy components only but the backup fails....here is the exact error from activity monitor;


18/08/2010 11:31:15 AM - Warning bpbrm(pid=13460) from client client_server_name: WRN - can't open object: Shadow Copy Components:\User Data\Distributed File System Replication\DfsrReplicatedFolders\IISConfig (BEDS 0xE000FEDD: A failure occurred accessing the object list.)
18/08/2010 11:31:16 AM - Error bpbrm(pid=13460) from client client_server_name: ERR - Unable to backup System State or Shadow Copy. Please check the state of VSS and associated Writers.INF - EXIT STATUS 69: invalid filelist specification
18/08/2010 11:31:17 AM - Error bpbrm(pid=13460) could not send server status message      
18/08/2010 11:31:17 AM - end writing; write time: 00:32:41
file read failed(13)



 I have checked the vss writers state and providers. This is ok.

16 REPLIES 16

J_H_Is_gone
Level 6
Tech doc  340189 (Etrack 1940549)

http://seer.entsupport.symantec.com/docs/340189.htm

Gerald_W__Gitau
Level 6
Certified
Yes, and the fix has been applied. The concerning factor is that no backup takes place just fails with error 13. On the article it says backups complete with status 0 even though the DFSR data has been skipped.

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Wild guess, but knowing you you have probably already had a look at this:

http://seer.entsupport.symantec.com/docs/354674.htm

It also doesn't explain why it's working when you backup C:\.

Maybe you've discovered another bug?
Do you have a bpbkar log?
 

Gerald_W__Gitau
Level 6
Certified
Yes, I checked that too...I have mailed you the bpbkar log...I logged a call with Symantec yesterday as critical and had to re-assign the ticket yesterday 1 pm CAT but still no word from them. So it looks like I will downgrade the client to 6.5.4 or 6.5.5. I know 6.5.4 works in windows 2003 not sure on windows 2008......

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified
Each file name in bpbkar is enumerated, but all of them fail with:

WRN - can't open file: Shadow Copy Components:\User Data\Distributed File System Replication\DfsrReplicatedFolders\xxx\xxx (WIN32 3: Unknown error)


Before this, there seems to be a snapshot problem as well. Do you have a bpfis log?

Seems there's no knowledge of a prior full backup and that the sched is Incr (not sure if this is causing the problem...Do you get the same result when running a Full?):

12:56:59.682 PM: [3192.4316] <4> SnapshotAssociation::AdjustUnboundedAssocations: INF - QueryDiffAreasForVolume returned failure status; no Diff areas for this volume
12:57:02.334 PM: [3192.4316] <16> VssSnapshot::V_VssVolumeSnapshot(): INF - VSS Volume Set prepare failed

.....

12:57:02.366 PM: [3192.4316] <4> dos_backup::V_InitializeShadowCopy: INF - ============================================
01:30:33.284 PM: [3192.4316] <2> ov_log::V_GlobalLog: ERR - BEDS_AttachToDLE():FS_AttachToDLE() DeviceName:'System?State' BackupReason:0x400 Failed! (0xE000FEC9:A failure occurred accessing the Writer metadata.



W2008 support listed on p. 27 of Compatibility Guide.

Hope Symantec Support contacts you SOON! Maybe wait with downgrade until they're happy that they have all the logs that they need.

Gerald_W__Gitau
Level 6
Certified
bpfis does not log anything when doing shadow copy components.....but it does when doing normal backup.

The full backup also fails.

I will downgrade to 6.5.6 tonight and let you know how it went. Thanks for your help.

Gerald_W__Gitau
Level 6
Certified
Ok..I downgraded the client to 6.5.6, but still the backup failed exactly like 7.0. So I went a step down to 6.5.5 and now I get a status 1 error. Much better but still all shadow copy components not backing up.

I managed to call Symantec and refused to hangup till I spoke to a Techie, we have been on webex for the last 3 hrs and still no backup. Took a break and will continue troubleshooting again. I hope to have this resolved before close of business today...

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified
Hi Gerald

Any progress on this? Hopefully working backups by now?

tbl
Level 3
Partner Certified

Hello,

have you had any information about symantec ?

I have the same problem with a client in 6.5.4 and servers in 6.5.5

the backup stops when processing the directive Shadow Copy Components.

 

thomas

KEELIN_HART_2
Level 4

Hi Gerald / Thomas. Did you find a resolution? We are having similar issue with NBU 6.5.4 server and client.

Thanks.

J_H_Is_gone
Level 6

I am using 7.0.1 and DFSR and 2008 servers.

The change between 6.5.x and 7.0.x is that now the DFS drive data is going to try to backup inside the SCC job.

So it will be a big job and always a full.

Example.

I have drive E as a dfsr drive.

I do all local drives.  E only gets a few meg.

SCC is huge - you need to make sure that it is using VSS (client attributes in master server properties).  SCC does not do incrementals.

So if you get the SCC job to finish and you go look at it in the BAR you will see DFS and all your files that reside on your DFS drives.

My issue was that it took 2.5 days to backup the DFS drives in the one BIG SCC job.

So I turn dfs off (via bpstart_notify and bpend_notify) so the backup of the E drive gets those files ( and is able to do an incremental) and the SCC does all its other stuff but does not backup the files from the DFS drives.

Giroevolver
Level 6

Rather than stop the DFS service you could just disable the VSS writer and always do a flat file backup of DFS

http://www.symantec.com/business/support/index?page=content&id=TECH126944

Rajesh_s1
Level 6
Certified

Hi ,

We also faced lot of system state backup issues with respect to win 2008. In allmost all the cases we got it resolved .

We need to verify lot of things with respect to this , below are some of them which may help.

1. Verify the VSS writers in the client i.e.by typing  vssadmin list writers in the command mode . All the writers should be stable , sometimes it will be in waiting for completion state . EVen that is also ok . If you found any failures in the writers then system state backup will fail.

2. Then run the system state backup from windows native backup. If its works fine here then definately it will work in Netbackup.

3. verify the shadow copy space give for C:\ drive . It should be atleast 1Gb space else the backup will fail. It is better to increase some 2 - 4 GB .

4. In the NetBackup side verify the timeout setting of the client , by default it will be 300 sec but for win-2k8 system state backup it is not sufficent . Increase this by 1600 or more . 

5. In some of the SAP clustered servers you may face system state backup issue with passive nodes , even in Native windows backup also it won't work . This is because SAP servies are installed in cluster drives and that service will not be availabe in passive node. To resolve this issue you need to exclude some of the sap services from system writers.

Gerald_W__Gitau
Level 6
Certified

Just an update......

This issue was resolved and the actual problem was with Equallogic VSS. When the servers were installed, the HIT kit used was old and the MPIO was not properly reqistered on the servers. Everytime the backup was attempted this error was generated:

Log Name:      Application
Source:        EqualLogic
Date:          2010/08/21 08:46:02 PM
Event ID:      4002
Task Category: VSS
Level:         Error
Keywords:      Classic
User:          N/A
Computer:      xxxxxxxxxx
Description:
Error 0x80070490 opening SNMP connection to \\.\PHYSICALDRIVE1(0-8a0906-ca190ec06-928000000414c47e).

 

Log Name:      System
Source:        MSiSCSI
Date:          2010/08/21 08:46:02 PM
Event ID:      113
Task Category: None
Level:         Warning
Keywords:      Classic
User:          N/A
Computer:      xxxxxxxxxxx
Description:
iSCSI discovery via SendTargets failed with error code 0xefff0024 to target portal *10.20.1.50 0003260 EBDRV\L4SC&PCI_164F14E4&SUBSYS_112314E4&REV_00\5&8571ec8&0&30050300_0 .
 


 

loori
Level 5

We encounter the same problem using 7.01 (Master Solaris 10 Sparc, Media Solarisx86 AMD, Client W2k8R2), RC is 50, message text the same (ERR - Unable to backup System State or Shadow Copy. Please check the state of VSS and associated Writers.ERR - Unable to backup System State or Shadow Copy. Please check the state of VSS and associated Writers.ERR - Unable to backup System State or Shadow Copy. Please check the state of VSS and associated Writers.ERR - Unable to backup System State or Shadow Copy. Please check the state of VSS and associated Writers.INF - Estimate:-1 -1).

Not all the servers have that, on some the backup is o.k. - but not in ShadowCopyComponents, but normal files, some in ShadowCopyComponents, some in normal files and ShadowCopyComponents hangs.

The servers are all setup similarly and the policy are similar.

I'll try to address that at the next usergroup meeting in Munich

Conclusion from usergroup meeting: Everybody faces problems backing up DFS-R . . .

We changed the policies to flashbackup, seems to work.

Btw, there is also a problem restoring DFS backup from SCC due to the path. We always have to move the restored files manually, they never restore correctly - no matter if via "restore everything to . . ." or change "selected destination . . ."

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Hi Gerald

I know it's a l-o-n-g time down the line... How exactly was the problem resolved?

Did you update/upgrade Equallogic VSS or remove it?