NetBackup 6.5 MP1 - NDMP

[Dirk_Mueller]

HI@ll,

I try to establish NDMP-Backups for our test-installation of NB 6.5 MP1.

I think I have installed it like our Prod-installation 6.0 MP5.

I always get error 99 and the info:

NDMP_LOG_ERROR 0 Write block size of 64 is not a 512 bytes mutiple

Here are our settings:

SIZE_DATA_BUFFERS: 262144

NUMBER_DATA_BUFFERS: 64

this is the same in both installations.

I don´t know the reason for this error-messages.

Kind regards

Dirk

[zippy]

Dirk,

Can your master communicate with your NDMP server?

Do you have a NDMP license on your master?

Do you have the NDMP PDF in downloaded? 

Do you have NDMP running on any other system?

[Omar_Villa]

What kind of NDMP configuration you have? Standalone NDMP, 3-Way NDMP or Remote NDMP, then have you test the user and password to connect to your filer? if you dont know what Im talking about, check this:

http://ftp.support.veritas.com/pub/support/products/NetBackup_Enterprise_Server/279269.pdf

hope this helps.

regards

[Dirk_Mueller]

Hi James,

my communication between master and NDMP server is ok.

The license on my master/media is ok

My NDMP-Backups in the prod-installation run without any problems - with the same vdm.

Kind regards

Dirk

[Dirk_Mueller]

Hi Omar,

I use remote NDMP.

The connection between master/media-server and my ndmp-filer is ok.

I think I have installed everything as in our prod-installation.

[Casey_King]

Did the initial backup work after setting configuring the NDMP backup?  Have there been any changes to the data.  The 99 means:

1. There have been no change to the data since the last backup

2. You surpassed a threshold on drive space which doesn't allow the checkpoints to be created.  Disk space utilization must not surpass 90% or you will run into this problem.

Hope this helps.

[Omar_Villa]

How are the NDMP-Drives that you assign? are they healthy? match the tpconfig, tpautoconf and vmglob drives serial number, path and name, also if you can enable the ndmpagent log at level 5 maybe you can get something there, here is a cool pdf to troubleshoot NDMP issue.

http://seer.entsupport.symantec.com/docs/287646.htm

hope this helps.

regards

[Casey_King]

I am not sure having a discussion about the drives is going to resolve this problem.  Maybe I am reading it wrong, but the status code 99 is a space utilization or no data change.  It is more of an informational status code not an error message.  From what I am reading here, the backup kicks off, assigns drive and media, mounts, but then ends because there is nothing to backup or the checkpoint isn't created.  This is where the issue most likely resides.

The PDF you posted was nice and I saved it because it gives a good overview of the process taken to do the backup.  Thanks for posting it.

[Dirk_Mueller]

Hi, my drive is healthy and I think it is correct installed.

I have updated the loglevel - ndmpagent, ndmp, global logging level.
Here´s a part of the ndmpagent-Log:

ndmp_data_start_backup returned 0 (NDMP_NO_ERR)
10:17:38.562 PID:3748 state change from STARTING to ACTIVE
10:17:39.078 PID:3748 EmEventManager: input on socket 1836
10:17:39.078 PID:3748 InputHandler entered - session = 00808140
10:17:39.078 PID:3748 [Error] NDMP_LOG_ERROR 0 Write block size of 64 is not a 512 bytes multiple
10:17:39.250 PID:3748 EmEventManager: input on socket 1836
10:17:39.250 PID:3748 InputHandler entered - session = 00808140
10:17:39.250 PID:3748 received DATA_HALTED reason = 3 (NDMP_DATA_HALT_INTERNAL_ERROR)
10:17:39.250 PID:3748 halt reason = THIS OPERATION FAILED
10:17:39.250 PID:3748 state change from ACTIVE to HALTING
10:17:39.250 PID:3748 received MOVER_HALTED reason = 1 (NDMP_MOVER_HALT_CONNECT_CLOSED)
10:17:39.250 PID:3748 Records 0 KbytesThisPath 0 KbytesThisFragment 0
10:17:39.250 PID:3748 [Error] NDMP backup failed, path = /root_vdm_14/FS_VDM_14_bckckpt
10:17:39.281 PID:3748 data state = NDMP_DATA_STATE_HALTED
10:17:39.281 PID:3748 ndmp_data_stop status = 0
10:17:39.281 PID:3748 mover state = NDMP_MOVER_STATE_HALTED
10:17:39.281 PID:3748 ndmp_mover_stop status = 0
10:17:39.281 PID:3748 state change from HALTING to IDLE

ndmp-log:

os_type="DartOS"
10:14:35.950 PID:3512 b170a8 .   os_vers="EMC Celerra File Server.T.5.5.29.1"
10:14:35.950 PID:3512 b170a8 .   hostid="abc1997"
10:14:35.950 PID:3512 b170a8 **  End NDMP_CONFIG_GET_HOST_INFO
10:14:35.950 PID:3512 b170a8 ->  (612) 5 [0] 10:14:35 NDMP_TAPE_OPEN (0x300) NDMP_NO_ERR (0x0)
10:14:35.950 PID:3512 b170a8 .   device="c112t0l3"
10:14:35.950 PID:3512 b170a8 .   mode=NDMP_TAPE_READ_MODE
10:14:35.950 PID:3512 b170a8 **  End NDMP_TAPE_OPEN
10:14:36.043 PID:3512 b170a8 <-  (612) 7 [5] 10:14:36 NDMP_TAPE_OPEN (0x300) NDMP_NO_ERR (0x0)
10:14:36.043 PID:3512 b170a8 .   error=NDMP_NO_TAPE_LOADED_ERR (0xa)
10:14:36.043 PID:3512 b170a8 **  End NDMP_TAPE_OPEN

bptm-log:

io_open: report_attr, fl1 0x00000049, fl2 0x00000000
10:17:35.078 [3956.3680] <2> io_open: file E:\VERITAS\NetBackup\db\media\tpreq\drive_IBM.ULTRIUM-TD2.001 successfully opened (mode 2)
10:17:35.078 [3956.3680] <2> write_backup: media id 000011 mounted on drive index 0, drivepath c112t0l3, drivename IBM.ULTRIUM-TD2.001, copy 1
10:17:35.078 [3956.3680] <2> io_read_media_header: drive index 0, reading media header, buflen = 65536, buff = 0x11ddff8, copy 1
10:17:35.078 [3956.3680] <2> io_ioctl: command (5)MTREW 1 from (bptm.c.7774) on drive index 0
10:17:35.312 [3956.3680] <2> io_read_media_header: ndmp_tape_read_func returned 65536
10:17:35.312 [3956.3680] <2> io_read_media_header: block read is not a NetBackup or Backup Exec media header, len = 65536, media id 000011, drive index 0, data is unknown
10:17:35.312 [3956.3680] <2> io_write_media_header: drive index 0, writing media header
10:17:35.312 [3956.3680] <2> io_ioctl: command (5)MTREW 1 from (bptm.c.9019) on drive index 0
10:17:35.391 [3956.3680] <2> io_write_block: ndmp_tape_write_func returned 1024
10:17:35.391 [3956.3680] <2> io_ioctl: command (0)MTWEOF 1 from (bptm.c.9060) on drive index 0
10:17:38.406 [3956.3680] <2> io_write_media_header: report_density, 0x42 [LTO-CVE]

I think I must adjust the Write block size, but I don´s know where I can do that.It´s curious, because in our prod-installation I can find the value 65536 in the file SIZE_DATA_BUFFERS_64 - and it work´s fine.

All tapes and drives are the same in all robots.

Any idea???

Kind regards

Dirk

[Andy_Welburn]

Have a search in the knowledge base for SIZE_DATA_BUFFERS

( http://www.symantec.com/business/support/overview.jsp?pid=15143 )

Just had a quick browse myself & there are a LOT of documents there with recommendations & caveats & hopefully some will relate to your set-up/architecture.

[Yashwanth · Accepted solution]

I used to get this error. But in our case the culprit was the backup path. Backup Path should be /Vol/volumename/. In my case I changed the path under backup selections and everything worked fine. Give it a try.