NBU 7.1 / GRT AD / Browse timeout with a big OU (most 3000 objects)

[Stephane_COLIN]

Hi,

I have a GRT AD restoration problem with a big OU (most 3000 objects).

My context :

   - 1 Master server w2k3 x86 with NetBackup 7.1 (dedicated Master)

   - 1 Media server w2k3 x86 with NetBackup 7.1 + STU "basic disk" + NFS Services configured (RPC External Data Representation / RPC Port Mapper)

   - 1 Client w2k3 x86 with NetBackup 7.1 AD 2003 natif + NFS Services configured (Client NFS + MS Services of NFS ADM + RPC External Data Representation + RPC Port mapper)

      * AD with a lot of OU and 1 big OU (most 3000 objects)

Backup Status 0 (no error/warnig) :

17/05/2011 19:00:00 - Info nbjm(pid=18516) starting backup job (jobid=818094) for client clntad.dom.com, policy pl_grt_AD, schedule full 
17/05/2011 19:00:00 - Info nbjm(pid=18516) requesting STANDARD_RESOURCE resources from RB for backup job (jobid=818094, request id:{25D32B66-685E-4C1A-8130-D0B95E97F240}) 
17/05/2011 19:00:00 - requesting resource MES-03P-STU-DISK
17/05/2011 19:00:00 - requesting resource masternbu.NBU_CLIENT.MAXJOBS.clntad.dom.com
17/05/2011 19:00:00 - requesting resource masternbu.NBU_POLICY.MAXJOBS.pl_grt_AD
17/05/2011 19:00:00 - granted resource masternbu.NBU_CLIENT.MAXJOBS.clntad.dom.com
17/05/2011 19:00:00 - granted resource masternbu.NBU_POLICY.MAXJOBS.pl_mes_grt_AD
17/05/2011 19:00:00 - granted resource MediaID=@aaaaf;Path=D:\MES-03P-STU-DISK;MediaServer=medianbu
17/05/2011 19:00:00 - granted resource MES-03P-STU-DISK
17/05/2011 19:00:00 - estimated 1900806 Kbytes needed
17/05/2011 19:00:00 - Info nbjm(pid=18516) started backup job for client clntad.dom.com, policy pl_grt_AD, schedule full on storage unit MES-03P-STU-DISK
17/05/2011 19:00:02 - started process bpbrm (5800)
17/05/2011 19:00:07 - connecting
17/05/2011 19:00:09 - connected; connect time: 00:00:02
17/05/2011 19:00:15 - begin writing
17/05/2011 19:13:40 - end writing; write time: 00:13:25
the requested operation was successfully completed(0)

Browse and restore GRT is OK : 1 object on a small/normal OU (less 100 objects)

Browse is not OK : big OU (most 3000 objects)

Tests / Test Browse from Master SERVER  / from Media SERVER / from client

   - NetBackup starting a restore job (without information)

   - nbfsd and bpdm starting on media server

      * bpdm read disk backup (Analysed with processmon.exe and with bpdm logs)

      * nbfsd logs an activity with VNETD (on local and with Master NBU) (analysed with processmon.exe and with nbfsd logs)

      * a temp.edb is create on <NBU_INSTALL_PATH>\NetBAckup\Temp\...

   - Browse end with file read failed (restore job finish with status 0)

Have you any idea ?

Stéphane COLIN

C-Storage

[Stephane_COLIN]

I tested another scenario --> directly backup AD to master server (on basic disk)

   - 1 Master/Media server w2k3 x86 with NetBackup 7.1 + STU "basic disk" + NFS Services configured (RPC External Data Representation / RPC Port Mapper)

   - 1 Client w2k3 x86 with NetBackup 7.1 AD 2003 natif + NFS Services configured (Client NFS + MS Services of NFS ADM + RPC External Data Representation + RPC Port mapper)

Now, the Browse and restore operations work fine.

Probably a network problem (netbackup side, operating system side, hardware side, ... i don't know).

I continu my analisis, and will keep you informed

Stéphane COLIN
C-Storage

[Stephane_COLIN · Accepted solution]

After had disabled all "Offload" and "flow control" driver parameters on NIC (Master and Media), browse and restore operations are OK.

Stéphane COLIN

C-Storage