Filer-to-Server NDMPV4

[hkrupp]

Hi, I try to setup Filer-to-Server NDMP-Backup between TSM 5.4 Server and NetApp ONTAP 7.2.2. I activated ndmpd on the Netapp and followed the small section in the chapter 7 (p. 159 of the 5.4 Admin Guide. I registered a node and defined a datamover like in the documentation. Then I try to start the Backup and always get:

ANR2017I Administrator ADMIN issued command: BACKUP NODE netapp1 /vol/vol0
ANR0984I Process 5 for BACKUP NAS (FULL) started in the BACKGROUND at 11:32:57 AM.
ANR1063I Full backup of NAS node NETAPP1, file system /vol/vol0, started as process 5 by administrator ADMIN.
ANR4794E The NAS file server netapp1 failed to open an NDMP data
connection to the TSM tape server. Please verify that the file server is capable of outbound data connections.
ANR4728E Server connection to file server NETAPP1 failed. Please check the attributes of the file server specified during definition of the datamover.
ANR1096E NAS Backup to TSM Storage process 5 terminated - storage media inaccessible.
ANR0985I Process 5 for BACKUP NAS (FULL) running in the BACKGROUND completed with completion state FAILURE at 11:32:57 AM.

I can see a Filespace being created after the definition of the datamover and the connection between the TSM-Server and the NetApp seems to work. After a "ndmpd debug 50" on the Netapp I get the following:

...
Jul 11 11:32:57 CEST [ndmpd:36]: Data connection type: 1
Jul 11 11:32:57 CEST [ndmpd:36]: TCP Addr: 0.0.0.0
Jul 11 11:32:57 CEST [ndmpd:36]: TCP Port: 0
Jul 11 11:32:57 CEST [ndmpd:36]: Error code: NDMP_CONNECT_ERR
Jul 11 11:32:57 CEST [ndmpd:36]: IOException: Connection refused
Jul 11 11:32:57 CEST [ndmpd:36]: NDMP message type: NDMP_DATA_ABORT
Jul 11 11:32:57 CEST [ndmpd:36]: NDMP message replysequence: 8
Jul 11 11:32:57 CEST [ndmpd:36]: Message Header:
Jul 11 11:32:57 CEST [ndmpd:36]: Sequence 0
Jul 11 11:32:57 CEST [ndmpd:36]: Timestamp 0
Jul 11 11:32:57 CEST [ndmpd:36]: Msgtype 1
Jul 11 11:32:57 CEST [ndmpd:36]: Method 1027
Jul 11 11:32:57 CEST [ndmpd:36]: ReplySequence 8
Jul 11 11:32:57 CEST [ndmpd:36]: Error NDMP_NO_ERR
Jul 11 11:32:57 CEST [ndmpd:36]: Error code: NDMP_ILLEGAL_STATE_ERR
Jul 11 11:32:57 CEST [ndmpd:36]: NDMP message type: NDMP_DATA_STOP
Jul 11 11:32:57 CEST [ndmpd:36]: NDMP message replysequence: 9
...

Seems to me that the NetApp is instructed to connect to 0.0.0.0 Port 0. I think there should be the IP of the TSM-Server and a Port defined by the TSM-Server.

Any hints?

Thank you,

Heiko.

 

[chad_small]

does the datamover have the root password for the NetApp? Also check that the TSM drives are defined on the NetApp and you have them pathed to the TSM server. You do have the TSM tape drives zoned on the NetApp right??

 

[hkrupp]

Hi,

the TSM Server has the root password. The communication between TSM-Server and Filer is ok. After the creation of the datamover I got a WAFL-Filespace with the correct size.

I did not define any tapedrives because I wanted to use the new Filer to Backupserver-Feature of TSM 5.4.0 (NDMP V4) where the Controlstream and the Datastream goes over the Network to the TSM Server and the Dump and TOC is written into the normal NATIVE Storagepools (TSM 5.4 Adminguide Chapter 7). This is a very new feature only slightly covered in the Manual and no Redbook or other config-examples could be found.

Thank you,

Heiko.

 

[hkrupp]

Hi,

the problem is solved. TSM uses gethostbyname to get its IP-Address. Unfortunately I had an old entry in my /etc/hosts and so the Server got 127.0.0.2 (test on an unused interface) as its IP. The NetApp didn't accept this IP and used 0.0.0.0 instead. After I corrected this I could find the correct IP in the TSM Trace and in the NetApp debug-log. Now everything works fine

Heiko.